From Dr. Roy Spencer’s Global Warming Blog
Roy W. Spencer, Ph. D.
SUMMARY
- My analysis of UK daily high (Tmax) and low (Tmin) temperatures during 1960-2025 using a station relative bias removal method produces UK-average summer temperature variations essentially identical to the very different UKMO methodology.
- In both my and the UKMO analysis, 1976 (not 2025) was the hottest summer in daily high temperatures (Tmax), with 2025 taking 3rd or 4th place; the “record” hot year of 2025 was due to nightly low temperatures (Tmin) being anomalously warm.
- The average of the three hottest daytime temperatures in each summer month put the summer of 2025 in 4th place since 1960, behind 1976, 1995, and 2022 (which were essentially identical).
There has been criticism of the UK Met Office’s methodology for monitoring long-term changes in UK-average temperatures, starting with Tallbloke’s (Ray Sanders’) blog post on 31 October 2024. A major criticism that Tallbloke has is the fact that most UK stations do not meet the World Meteorological Organization (WMO) criteria for a good climate monitoring station. The UKMO doesn’t actually use the WMO quality classification system, but their own 4-tiered system. Another criticism is that many UK stations have closed in recent years, and so those stations are, in effect, estimated (“fabricated”?) from surrounding stations.
No Station is Perfect
On the subject of which WMO (or UKMO) class of station is suitable for long-term climate monitoring, I think it’s important to note that a station could be placed in a non-natural, anomalously warm urban environment, but as long as that environment stays the same over time, it can probably still be used for climate change monitoring.
For example, the urban heat island (UHI) effect of London was described over 200 years ago by Luke Howard. Even if London is significantly warmer than the surrounding rural areas, it might be that there has been little additional UHI warming since then, and so a downtown London weather station might be adequate for monitoring large-scale climate change, since I have no reason to believe that (say) 1 deg. C of large-scale warming will lead to city warming substantially different from 1 deg. C.
On the additional subject of replacing a closed station with estimates from surrounding stations (which NOAA also does because so many of their UNHCN stations in the U.S. have closed, a process that has also been criticized), I believe it is a little disingenuous to claim those data are “fabricated”. Rather than continuing the closed station record with estimates from surrounding stations, one could just use the surrounding stations, which is the same thing.
The UKMO’s Fancy High-Resolution Mapping of UK Temperatures
The Met Office divides the UK land mass into tiny (1×1 km) grid cells, and the temperature in each one is estimated from the nearest station(s) using average, regression-based adjustments for elevation, latitude, longitude, terrain shape, coastal proximity, and land use variations. The result is a seemingly complete coverage of UK for the purpose of temperature monitoring:
And I get why this is done: the UKMO primary mission is to provide daily weather monitoring and forecasts, and given limited station data providing actual measurements, their system provides useful temperature estimates in areas far removed from actual weather stations.
Of course, all of this high-resolution fanciness must be anchored by actual measurements, and in the daily Global Historical Climate Network (GHCNd) database, only ~100 stations exist across the UK in recent years. (There were very few GHCNd stations before 1960, so I will address temperature change only since then here). This means only 1 in ~2,400 UK grid cells has an actual temperature monitoring station in the GHCNd dataset, which is the dataset all global temperature monitoring efforts rely upon. While the UKMO might have access to somewhat more stations than are included in the GHCNd dataset, my point will remain valid.
Nevertheless, this doesn’t mean that long-term climate change can’t be monitored with the existing station network. What complicates matters is that stations come and go over time, and this can introduce biases that change over time and corrupt long-term estimates of temperature change. How one accounts for, and adjusts for, these changes is not a settled matter.
Removing Relative Biases Between Stations
From what I’ve been able to glean, the UKMO does not actually calculate and remove relative biases between stations. Instead, they use the above-described strategy to evaluate how station temperatures vary with latitude, longitude, elevation, proximity to the coast, land use (e.g. urbanization), etc., then apply regression-based techniques to estimate temperatures on the 1×1 km grid. This has no doubt involved considerable effort, and having done similar kinds of data analysis myself, it’s a complex task.
A simpler way of monitoring climate change is to assume that long-term (in the current example, 65 years) warming trends that actually exist in nature are pretty uniform across the UK. If this assumption holds, we can just take whatever stations exist over time, no matter where they are located or what their local microclimate-induced biases are, and quantify how the temperatures at each one varies over time, and then average all of those variations together. This methodology is somewhat similar to that of Hansen and Lebedeff, 1987, as well as our UAH satellite global temperature dataset.
In my implementation of this relative bias removal method, I start with the stations having the longest periods of record. In the UK, only 3 stations have had continuous records in all 126 years from 1900 to 2025: CET Central England, Armagh, and Stornoway Airport. (Only 31% of the UK stations had periods of record at least half as long, 63+ years). I average those 3 stations together. Then, I take the station(s) with the next-longest record (Oxford, 124 years), compute the average difference with the original series, and add it to the series to make a new 4-station average. This is done sequentially for all (148) stations in the UK since 1900 that have at least 2 years of record, going down the list from the longest periods of record to the shortest. Again, since there were few stations before 1960, the following plots cover variations since 1960.
Amazingly, the result using this simple relative bias removal method produces yearly summer-average temperatures (average of daily high [Tmax] and daily low [Tmin]) that are nearly identical to the much fancier UKMO methodology:
In this plot (as well as the others, below) for display purposes I have removed a small (~1-2 deg C) temperature offset due to my use of the original 3 stations for an absolute temperature baseline, whereas the UKMO uses their gridded estimate of the entire area of the UK. The linear trends in the above plot are essentially identical, at +0.27 C/decade.
But that record high did not exist for the daytime high temperatures. As seen in the next plot, 2025 was very similar to several years since 1995, while 1976 holds the record for hottest summertime daily high temperatures:
So, where did the 2025 record come from? It was due to the nighttime temperatures being so warm (although I don’t see how 53 deg. F is is insufferably warm). This was true in both analyses of the station data:
Finally, since I am analyzing daily temperature data, I can compute the average of the three hottest daytime temperatures in each summer month, which produces this:
For this statistic we see that the record is a 3-way tie between 1976, 1995, and 2022. We also see a stronger warming trend (+0.40 C/decade vs ~+0.26 C/decade in all-days average Tmax and Tmin). I suspect this is due to more Saharan air intrusions in recent decades, which are the primary cause of excessively hot days in the UK.
Conclusions
Despite criticisms of the UKMO data and methods for computing UK-average temperatures, I find that a simple bias-removal method of combining all available UK stations produces essentially identical results to the much more complex UKMO methodology. It should provide some vindication for the UKMO methodology in the context of climate temperature trend monitoring.
The record hot summer of 2025 in the UK was in the nightly minimum temperatures, not in the daytime maximum temperatures. This is true in both my analysis and that of the UKMO.
Finally, neither my nor the UKMO method accounts for possible changes in stations over time, such as an increasing urban heat island (UHI) effect at some stations. Based upon our work on this in recent years I suspect this effect since 1960 would be small, but I don’t know that for sure.
Many thanks to Dr Spencer…there is only a snowball’s chance in Hades of getting real weather forecasting in the UK…for various reasons…
Roy’s method is OK, in terms of inter-station bias, but unlike the UKMO method, it does not do spatial weighting. That doesn’t matter so much for a small region like the UK, where they stations they have chosen are reasonably well distributed. I wouldn’t call it simple, though.
The island of Great Britain is often described as a ‘small island’ especially by the Brits themselves. However it is the 9th largest globally, coming between Canada’s Victoria and Ellesmere Islands. It is over 600 miles north to south on the main island and 800 if you include islands. 600 miles is roughly the distance between Sydney and Gold Coast if the Internet is to be believed.
What counts is size relative to climate variation. And when England has a warm summer, usually Scotland does too, etc.
You’ve obviously never lived in Scotland
I have experienced two quite warm summers in Scotland (1975, 1999).
Holiday or lived there? If a holiday what time of year.
The period 1972-76 was exceptionally dry in Scotland with the record 3 month drought in the East in 1973. Not surpassed in 1976 or before or since.
Scotland escaped Michael Fish’s Hurricane but had one that England escaped in January 1968.
We can all throw personal experience into the discussions. But Scotland certainly doesn’t follow SE England in many things including weather.
Travels. 1975 was holiday in July (sleepless warm night in Inverness!). 1999 was a spell at Edinburgh University, also I think in July.
But my point is just that the weather goes together. If it’s relatively warm in England, it is likely to be relatively warm in Scotland.
As it was summer you probably never heard anyone say “Oh no there’s been snow in London, how are the poor souls going to cope with 3 flakes”
But quite frequently what used to be known as a ‘ridge of high pressure extending north from the Azores’ only extended to the southern half of Great Britain whilst the normal South Westerly airstream brought dreich weather to Scotland. Equally cold coming from the arctic often didn’t get past Scotland.
1976 must have been quite warm in Scotland.
It was. The streets in Ayrshire were dusty, like the Mediterranean and the perennially green fields looked brown. I flew from Glasgow to the South of France with a window seat and saw no green below until I was over Provence!
And was that “warm” weather in Scotland unbearable?
Harold The Organic Chemist Says:
For an Edinburgh temperature check, I went to:
https://www.extremeweatherwatch.com/cities/edinburgh/average-temperature-by-year. The Tmax and Tmax data are displayed in a table from 1960 to 2024. Here is the temperature data for these two dates:
Year—–Tmax—–Tmin—–Tavg Temperatures are ° C
2024—–13.7——-6.6——-10.2
1960—–12.7——-5.3——– 9.3
Change-+1.0—–+1.4——–+1.2
Here is data for Aberdeen:
Year—–Tmax—–Tmin—–Tavg
2024—–12.8——-5.8——-9.3
1960—–11.7——-4.4——-8.1
Change-+1.2—–+1.4——+1.2
CO2 Concentrations:
2024: 425 ppmv (0.83 g CO2/cu. m. of air)
1960: 317 ppmv (0.63 g CO2/cu. m. of air)
The temperature data reveals that Scotland is a chilly country.
Note that although the concentration of CO2 in air increased by 33%, the increase had little effect on air temp. Note how little CO2 there is the air.
One question that needs be answered is: What is the natural variation in Tmax and Tmin. Note how little mass of CO2 the is in the air.
It would be of interest to get Tmax and Tmin data for a summer month such as July.
When you go to the above URL, you have the option of acquiring Tmax and Tmin data for each month of the year. There is a selection box for a month next to the Month label.
Be sure to go to home page: https://www.extremeweatherwatch.com On the home page, there are links in light blue to many sites around the world. To obtain temperature data for a country, replace “cities” with
“countries” in the URL. Enter the country name in lower case letters. If the country name is two words connect these with a hyphen.
The west coast of Scotland is milder than the east coast due to the Gulf Stream, so much so it’s milder than parts of England. Also the location of the jet stream also impacts on local weather conditions.
Correct, on the coast to the south of me at Rhinns of Galloway there are palm trees that survive winter and in fact flourish. I am to far inland and too high to be frost free so stick to native plants.
Data from nearby pristine sites shows the 1930s,40s on average warmer than the period from 2000-2020.
Population growth and densification is also basically linear, resulting in linear warming by urban heat effects.
Valentia Island Met Station in SW Ireland also shows the 1930’s as warmer than today.
That’s the station I was talking about.
One of the most consistent and pristine sites anywhere.
I ran a guest post by Stephen Connolly based on data from Valentia – excellent site and excellent data.
https://tallbloke.wordpress.com/2025/03/10/comparing-temperatures-past-and-present-some-quality-data-analysis-from-an-interesting-angle/
Excellent! May I ask how these figures compare with your own UAH data?
The measuring of the UHI influence might be made more difficult by not only the changing population in an urban locality, but also its spatial composition and density, horizontally and vertically. On top of this, changing energy “use per capita” figures, and where this energy is sourced before showing itself as waste heat might be hard to establish.
Combine this with changes in prevailing wind patterns possibly carrying warmed air either towards or away from thermometers, and only the most generalised speculations can be made.
However, if minima are rising more than maxima, and diurnal range is decreasing, an obvious explanation is increased anthropogenic heat, with affects on thermometers more apparent in the absence of the Sun.
There’s probably a PhD dissertation buried in there somewhere, one way or the other.
“Finally, neither my nor the UKMO method accounts for possible changes in stations over time, such as an increasing urban heat island (UHI) effect at some stations.”
Perhaps Roy should have read Ray Sanders analysis of the UK weather stations.
Half ot them don’t exist anymore and the data is made up, of those still in use the majority are below class 2 sites and have major siting problems..
Roy is talking about a serious dataset. None of the data is made up. It is all a record of thermometer readings.
Mr Stokes, you obviously know nothing about the UK weather stations to make such a comment.
Any dataset based on those weather stations is a joke.
If you read the article Grumps, your criticisms were actually addressed.
Like on the other thread.. ROFLMAO !!!
Sorry, but Met Office data is basically JUNK data, and shouldn’t be used for any decisions except if to hold a tea-party or not.
Just explain which thermometers are supplying the readings for Lowestoft, and where they are located. That’s all you have to do. What is so difficult about that?
I have done – see below. A thermometer in Lowestoft supplied GHCNd data until end 2004. Then the data finished.
If Roy is using any Met Office data, substantial amounts are just made up.
Sorry Nick is not aware of that fact.
That point was addressed.
Rubbish.. It is using Met data.. It is therefore using JUNK data.
Highly affected by bad sites, urban warming, airports, data manipulation and fabrication etc etc…
..and totally unfit for the purpose of long term “climate” study
I don’t know if Nick is correct. But if he is, what he is saying is that the Met Office data used for climate trend calculation is neither junk data nor made up data. It is real readings from real thermometers with no estimated readings in it.
If I have understood Roy Spencer’s post correctly he seems to be saying the same thing: that the stations, whatever their merits in absolute terms, are suitable for long term measurement of climate trends.
I don’t see it being an issue of such great importance for climate, because if I understand it correctly there are no great trends in any direction in the historical UK trends. Maybe I’m wrong about this, too?
Roy calculates UK trends of summer Tmin as 0.26C/decade and Tmax 0.29C/decade. I think that is right, and it is greater than the global all-month trend.
They seem to be reporting readings from Lowestoft, a town or city on the east coast of England. But there is no station in Lowestoft. So when this is pointed out they claim to be reporting a reading which is deduced from the readings of other nearby stations. But there are none, and they cannot seem to explain where these supposed other stations are, that are being used to estimate Lowestoft.
At least that is what I understand from Ray Sanders account. Is this correct (about the situation and about what Sanders says)?
If so, if we cannot say which station, if indeed any, a reading comes from, I don’t see how we can say that its a record of thermometer readings. If you cannot say which thermometer you are citing readings from, and if there is no reason to think there is one anywhere near the place whose temp you claim to be recording, is it surprising that people conclude that your readings are made up?
It should be pretty simple, there is an instrument at a given address, so at least name the place it is at. I would have thought, give the street, or the postal code. If they can’t or won’t do that, where is the temp reading coming from? How do we tell if it was correct? Compare it to another thermometer in the same place? Oh, wait….
“At least that is what I understand from Ray Sanders account. Is this correct (about the situation and about what Sanders says)?”
No! Roy is using GHCNd. You can find the file for Lowestoft at https://www1.ncdc.noaa.gov/pub/data/ghcn/daily/by_station/UKE00105899.csv.gz. It has daily entries, each from a thermometer reading at the Lowestoft station. The series finishes at end 2004. There is no substitution of readings from elsewhere, here or anywhere in GHCNd.
Ray is dishonestly conflating with a UKMO user app which allows you to interpolate to find met data at your location. This is a user service – it has nothing to do with the datasets used for UKMO national averages, GHCN or anything else.
So I was right about what Ray is saying, but what he is saying is totally wrong?
UKMO is not using estimated temperatures at all to calculate any UK climate trends? The estimated temps for (for instance) Lowestoft, are just basically weather forecasting, local use only, at best rough guidance, and not used for UK climate trend purposes?
It would be nice to have a reply from him on why he is making such an obviously erroneous claim if that is what he is doing.
“The estimated temps for (for instance) Lowestoft, are just basically weather forecasting”
Not even that. They are just for this interactive user utility. Not used for national averages or trends.
Anyway, there is nothing in that in GHCNd, which Roy is using. Easily verified by inspection.
[snip] They will probably delete me [yup, snip, edited for plain obnoxiousness ~ctm]
Mr Sanders, to add to your lack of integrity and potty mouth, you have a staggering superiority syndrome and judging from the above, a touch of “infamy infamy, they’ve all got it in for me”.
Just what qualifications do you have to disparage Mr Spencer and Nick Stokes as talking “utter crap”?
You know mine, though do not have the character to acknowledge it.
Nor to admit you do not even know what the WMO site criteria mean.
(The surrounding area should be representative of the landscape and NOT of it’s weather).
You are not even cognisant of the facts you are supposedly arguing for.
Roy Spencer Phd is talking “nonsensical drivel” eh?…
https://en.wikipedia.org/wiki/Roy_Spencer_(meteorologist)
Yes, of course he is, and you are Newton and Einstein rolled into one.
He did not “attack” you, he merely pointed out the obvious to those of us who see that your “arguments” are but diddly squat in your belief that the UKMO is perpetrating some sort of underhand campaign to push a warming world agenda.
As Nick and Roy have said, the UKMO is a national met service charged with providing data and forecasts to the UK general public and not to include everything they do into a climate database.
Your last sentence above actually beggars belief that it came from a grown man.
You have no answer.
OK, I now understand your and Banton’s point, and the argument that for climate trend purposes only actual recorded temperatures are being used and not the estimates which a user application is supplying. I suppose that the estimates can be useful in some circumstances.
This goes a long way to explaining something which my UK acquaintance has sometimes commented on: the temperatures they measure do not seem to correspond very well to those which the Met Office supplies for their location.
Presumably this is because the Met Office numbers are estimates based on stations which may be tens of miles away, whereas they are using a real thermometer in their back garden. No matter how careful they are with siting this is going to lead to discrepancies.
However the thing that is still troubling about the Met Office is the publicity about peak summer temps in the UK. It does seem to be true that they are feeding ridiculous and unrepresentative summer high temps into publicity, and this is in turn used, without denial, to imply that something strange dramatic and unprecedented is happening to UK weather.
Which I do not believe it is, its more or less linear with some large swings up and down, as in the famous Barbecue Summer. And so we have the ridiculous situation where summer temps well below those routine for Chicago or New York are publicized with heavy health and safety warnings, while on the news one can see footage of the beaches and parks full of oblivious sun bathers and picknickers.
I would add to this – I was in the UK in the recent supposedly record hot summers. The country was being told by Met Office and BBC how dangerous this was, to stay indoors with the curtains closed… etc.
Go outside, and it was pleasantly warm, including in the hottest part of the day. It was complete hysteria – in Charco’s sense: emotion out of all proportion to its object. It was probably not as warm as the locations to which Britons flock in their thousands in the Med every summer.
My impression is of a country that is perpetually on the verge of hysteria about something, and its not just the Met Office and BBC and weather or climate, you hear every day of some obsessive maniacal idea being pursued by some body, instead of their doing what they were actually set up to do. Its at the level of a UK curtain company suddenly deciding that its mission is to rescue the vole population of Slovakia. The NHS for instance, rather than looking after mothers in childbirth satisfactorily, spends its energies banning the term ‘mother’, and replacing it with ‘birthing parent’ – they are almost at the point of banning the term ‘woman’. Meanwhile maternity services in the UK are in a real crisis of quality of care.
On the present subject, the Met Office would be more credible if it simply said, we have no readings for Lowestoft because we closed the station. We have readings for xyz and they are these. And leave it to the public to figure out if they could use xyz as a guide.
Though what my acquaintance has found is that there you cannot substitute readings from some unknown place tens of miles away for your own thermometer in a back garden right where you live.
“Presumably this is because the Met Office numbers are estimates based on stations which may be tens of miles away, whereas they are using a real thermometer in their back garden. No matter how careful they are with siting this is going to lead to discrepancies.”
Yes, that is the key issue, although a common reason for discrepancy is less professional home reading. The issues are sampling and homogeneity. You want to say something about the weather of the country, and the only way is by sampling (finite number of thermometers). How to test – by seeing if you could predict station readings given nearby stations only. To the extent that you can, your samples are enough; inhomogeneity is what would require denser sampling.
If you can predict station readings, it follows that you should be able to predict between stations. Then you can put the whole story together by averaging or whatever. Predict doesn’t mean exact correspondence. It means a lack of bias, and not too much noise.
The real problem is what is the uncertainty of the prediction when using “nearby stations only”? The sampling is creating a random variable with a certain number of entries. What is the mean and better yet, the standard deviation (uncertainty) of that random variable. Microclimates vary considerably within short distances, even less than a mile. How is that taken into account when determining the uncertainty?
Would you allow a nuclear plant to average nearby radiation readings to determine the radiation at an unmonitored point? Would you allow a critical support plate that is subject to cracking to be used if the average hardness was calculated and the variance showed that some points did not meet the requirements? Why are temperature data critical to spending trillions of dollars any less important?
How is the uncertainty in the estimate dealt with when performing additional calculations?
Data should be treated as sacrosanct to the purpose it is used for. It should be declared fit for purpose only after critical review of its quality. Estimates only lead to the current replication problems plaguing the scientific community. Estimates derived from averaging hides critical information needed to judge its quality.
[snip] have posted the Talkshop’s response. https://tallbloke.wordpress.com/2025/12/21/mystifying-met-office-advocacy-defending-the-indefensible-and-disparaging-the-talkshop-not-on-my-watch/
I am frankly astonished why Mr Spencer felt the need to launch an attack on me. [snip]
The Central England Temperature series shows, excluding years before 1960, rated to 1 dp:-
Max+Min/2
2025=1976, 2018, 1995=2006=2022=2003
Min
2025, 2006, 2003, 2016=2021, 2022=2020
Max
1976, 2018, 2025, 1995, 2022, 2003=2006
I live in the south of the UK about 5 miles from the sea. Whenever it’s hot I place my digital thermometer in the same position in my garden, in the shade. Maximum temps occur around 4 in the afternoon. During hot periods the temp often exceeds 30C. But this summer the temps never reached 30C, which was actually quite unusual. Last summer it reached about 32C. So, yes, I don’t think it was anywhere close to a record as far as Tmax is concerned.
If it was a record Tmin I’m a bit surprised as I don’t remember any particularly hot or warm nights. Overall it was quite a pleasant summer, obviously a sign of the climate apocalypse!
I’m pretty sure that all these “records” have more to do with UHI than climate change.
I live in a similar place but nearer the coast. Before the 1990s exceeding 30C was unheard of. It happens nearly every year now.
It has been a crazy year weather wise, exceptionally consistent warmth especially at night. Yes, maximums have not been correspondingly record breaking. It’s almost certain now to be the warmest year in the entire CET series.
I was visiting London in July 1990.
Pretty sure there was a >30C spell, where the locals having the vapors.
If you are going to respond, at least read and understand the post first, or you just look silly.
oooohhhhkaaayyy . . .
(says I, backing away slowly, not making eye contact 🙁 )
Measured where.. We can basically ignore every Met site as being little more than junk data.
We’re you there for the Summer of 1976? It was over 30C for weeks.
I agree that day temps were never that high but I certainly agree with it being warm at night and I am in Surrey. Lots of nights needed a fan to get to sleep whereas normally it cools off quite a lot in the countryside.
” since I have no reason to believe that (say) 1 deg. C of large-scale warming will lead to city warming substantially different from 1 deg. C.”
This is simply wrong. You also have no reason to believe that a large-scale warming of 1 deg. C will lead to 1 deg. C warming in the city. It could easily be more in the city or it could be less.
This is the whole purpose of doing pristine temperature MEASUREMENTS (plural) over a wide spread area. If Spencer’s assumption is correct then all that is needed is to track ONE STATION on a global basis! That ONE STATION would be adequate to identify large-scale changes for the entire globe.
To obtain temperature data for a city, go to:
https://www.extremeweatherwatch.com/cities/city name
For city name, enter the name of the city in lower case letters. If the city name is two words, connect these with a hyphen. This URL format display all the weather and climate data for the city. At the end of page, there is a list of options for display of data. Most useful is:
Average-temperature-by year. In the “Select City” box, enter the city name starting with a capital letter. If the city’s temperature is in NOAA’s data base, the name will appear below the box. Click on the name to obtain the data.
To obtain data for a country use:
https://www.extremeweartherwatch.com/countries/country name/average-temperature-by year.
For the country name, enter the country name in lower case letters.
You can also obtain data for a state. Use:
https://www.extremeweatherwatch.com/states/state name/average-temperature-by-year.
You can use this cite to get data to determine if the temperature of surrounding area influences temperature in a city. I have check the temperature data for many cities and they all show recently warming due to the UHI effect. The exception is Adelaide in Oz which has no warming since 1887.
This summer it’s certainly been the warm nights that made the average temp so high rather then daily Max temps. Here in Scunthorpe the only time the Maximum temperature reached 30C or more was on June 30th (31.2C).
But we had a number of warm nights where the lows were 19C or higher.
June 22nd 19.1C
June 28th 20.1C
June 29th 19.4C
June 30th 19.3C
July 18th 19.6C
August 14th 19.6C
Even if London is significantly warmer than the surrounding rural areas, it might be that there has been little additional UHI warming since then
“Might be” isn’t the sane as “is.” There might be new construction that replaces an old building with a new one or a road or parking lot paved with material that artificially changes the temperature profile throughout the day. A new building could be taller or shorter, might block the wind or funnel more past the station, reflect or absorb more infrared radiation, vent more or less heat from an HVAC system into the area of the temperature gauge.
I don’t understand why gridding temperature data and assigning estimated numbers to each grid cell is a valid method. Please explain. I think it’s statistical nonsense that’s influenced more by the methodology than the measured data.
Why is that a better method than using a continuous record of Tmin and Tmax measurements at a single station and calculating a linear trend over time at that station for stations that are sited away from artificial heat sources that might bias the temperature? Instead of averaging data from numerous stations or making it up with gridded, estimated data, just produce a map of station data and linear trends at each station to see if temperature at some, most, or all stations is increasing over time. If CO2 is increasing downwelling radiation then it must be affecting every station on the globe, if not equally, close enough to equally to see a trend at every station.
There is no such thing as a global average temperature and a global average temperature anomaly. That’s just statistical gobbledygook that can be whatever you want depending on the calculations you use to derive it; no offense intended for your excellent work with satellite data. Let’s see linear trends at each station from actual measurements like the NOAA tide gauge map here:
https://tidesandcurrents.noaa.gov/sltrends
“I don’t understand why gridding temperature data and assigning estimated numbers to each grid cell is a valid method. Please explain. I think it’s statistical nonsense that’s influenced more by the methodology than the measured data.”
It isn’t a valid method. It’s a “convenient” method to make things easy. It simply doesn’t matter to climate science if its physically correct or not.
Hubbard/Lin, 2006, “Reexamination of instrument change effects in the U.S. HistoricalClimatology Network”
Climate science has chosen to ignore this study since 2006. It’s an inconvenient truth to them.
Thank you for pointing that out. I was being polite. It makes no sense to derive a global trend from fabricated data that has been artificially adjusted with algorithms that attempt to remove biasing factors, most of which are site-specific and not accurately removed unless that bias has been measured at the specific site and applied specifically to its data. It’s nonsense derived from nonsensical adjusted data. Averaging nonsensical data doesn’t improve its precision or accuracy. It’s still nonsense.
I love the concept of averaging temperature data on Pikes Peak with temperature data from Colorado Springs. Geographically close. Topographically vastly different. Or averaging data from the Southern Hemisphere with data from the Northern Hemisphere when they make up a multi-modal distribution. And anomalies don’t help because the difference in variance between cold temps and warm temps are inherited by the anomalies. Meaning they *still* represent a milti-modal distribution with different measurement uncertainties because of the different variances.
Climate science seems to be dominated by blackboard statisticians, not physical scientists. The belief that you can average intensive properties is colors the entire averaging heirarchy! I would love to give a climate scientist two rocks, one at 70F and one at 80F, and ask them if they are holding 150F in their hand!
These kinds of analysis are what needs to be done for all regions on the globe. Whether it is whole continents or smaller due to topography, the sum of the parts will tell a better story.
Having done years of forecasting in business my eyes tell me there are many “pauses” followed by short increases or decreases. Some pauses are short, and some are long. This is important in analyzing what is occurring.
I certainly wouldn’t forecast a continuous increase at the level that the linear regression or the exponential shows without investigating what external factors are causing the step changes. It certainly doesn’t fit with a continuous increase in CO2 levels.
Here is a random chart from a station in Peru. It is obvious that something changed early in this temperature record and simply relying on a linear regression would give a wrong impression. The shaded part is the standard deviation so one can see a statistical descriptor of the changing data.
As far as I can determine, passenger traffic grew from less than 500,000 in 2005 to over 1,500,000 in 2019
ie triple the passenger load.
Always be very suspicious of temperatures measured at airports. 😉
My point was to be very suspicious of Tavg. It is not a scientific value, it is a statistical parameter that hides information used in the calculations. You can see that Tmin would cause a rise in Tavg while Tmax remains pretty constant.
I KNOW most climate scientists, warmists, and reporters are not farmers. How do I know? You can plant 120 day maturing crops instead of 90 day. If Tmin rises from 30 degrees to 33 degrees you don’t need to go break ice for farm animals. People in rural Kansas are noticing this and question how the media claim that we are burning up.
Are people in the UK actually complaining about the heat? Probably all those folks from hot countries now living in mostly London probably like it.
While I have the highest regard for Dr. Spencer I am not convinced that the UKMO is off the hook. They are dishonest, you showed that yourself, why would they claim a record high year using only the Tmin? If we are going to show a temperature increase since preindustrial times I believe only actual temperature readings should be used and only at uncontaminated sites. Averages should not be used. Estimates should not be used. Absolute coverage is meaningless for this exercise because we don’t have absolute coverage now and damn sure didn’t have it in the 1800s. If you want to guess what the temperature will be tomorrow for a certain area then yes using readings from a nearby site is acceptable or maybe if you want to shade a map for demonstration purposes for what the temperature likely was yesterday. The UKMO has an agenda and whether they follow correct convention or not is meaningless if they misuse the data.
They dont. The calculation of daily temperature is based on Tmax + Tmin
So you can see that to “increase temperatures” you can either increase Tmax or increase Tmin or both and the point Roy is making is that increasing Tmin is a non-argument in terms of arguing for potential catastrophe resulting from “high” temperatures.
The Met Office have made adjustments to the CET monthly mean series, look at JJA differences in 1976, 1995, 2003, and especially in 2018:
https://www.metoffice.gov.uk/hadobs/hadcet/legacy/data/cetml1659on.dat
https://www.metoffice.gov.uk/hadobs/hadcet/data/meantemp_monthly_totals.txt
1976 being the depth of Globull Curling when eco-activist were crying impending agricultural doom from cooling since the 1930s.
Then a slow warming trend began.
Dr. Spencer ==> I’m afraid I can not agree with this: “ over 200 years ago by Luke Howard. Even if London is significantly warmer than the surrounding rural areas, it might be that there has been little additional UHI warming since then,”
1825 London consisted substantial government buildings of stone, mostly dirt or mud roads, wooden buildings, and even pastures, woodland and open spaces.
200 years later, land use has vastly changed, road are black asphalt, almost all buildings are concrete, stone and reflective glass … pumping out excessive heat in the summers and heating the interiors with a lot of loss of that heat into the environment. Cars and buses pump our heated gases day and night.
All of that adds up to UHI- additional UHI.
Kip, good to hear from you.
As more and more energy is used in all categories, waste heat is increased. It isn’t solely related to population growth. I see too many locations where Tmin is driving warming, even at USCRN stations.