Skip to content

Hottest Summer Evah? Not According To CET.

September 1, 2018
tags: ,

By Paul Homewood



The Met Office are desperate to show this summer as the hottest evah:



August was a fairly average month but it marked the end of what was a remarkable summer for many.

With one day to go, it is clear that 2018 has been one of the hottest summers on record for the UK, however, the margin between the mean temperatures at the top of the league tables (records dating back to 1910) is so small that at this point it is impossible to say if 2018 will be an outright winner.  It is very close to the record-breaking summers of 2006 (15.78C), 2003 (15.77C), and 1976 (15.77C) all of which are within 0.01C of each other.

The margin is so small that different datasets and different regions of the UK will have different ranking. Usually we will only quote statistics to the nearest 0.1C as differences smaller than this could result from small numerical differences arising from the statistical calculations. A more comprehensive analysis of the 2018 summer data will be undertaken early next week and data for summer 2018 will continue to be analysed over the coming months.


However, it looks more likely that it could be the warmest summer on record for England with the mean temperature standing at 17.2 C with one day to go, which would narrowly beat the record set in 1976 (17.0C). It is not going to be a record for Scotland, Northern Ireland or Wales where the records stand at 14.1C (2003), 15.5C (1995) and 16.1C (1995), respectively.

To get an even longer-term perspective our multi-century Central England Temperature* (CET) series dates back to 1659. In this dataset summer 2018 looks likely to slip behind the summers of 1976 and 1826. If we look back through the CET series only 10 summers recorded an average temperature above 17C. Six of those have occurred since 1976, and only two (1826, 1846) were pre 20th Century, which is consistent with the general picture of our warming climate globally and here in the UK.


Large questions remain about the accuracy of the Met Office’s calculations. They gloss over the CET, which is undoubtedly the gold standard of UK temperature series. In fact, this summer only ranks fifth warmest behind 1976, 1826, 1995 and 2003 (in that order). There is no comparison with 1976, which finished at 17.77C, compared to 17.27C this year.



In terms of average daily maximums, 1976 was 0.74C hotter than 2018.

It is hard to see how the Met Office can conclude that this summer is close to 1976 in the UK as a whole.

CET, of course, only covers Central England, but as the map shows, that is precisely where the intense heat has been this summer. Scotland, N Ireland and Wales are all nowhere being a record. Along with northern England, this should have pulled down the UK figures, in comparison with CET.


Whereas CET is very carefully homogenised and relies on three high quality stations, Rothamsted, Pershore and Stonyhurst, the UK numbers are derived from a large number of sites, some of which have dubious provenance.

Although the Met Office don’t actually publish the stations used (at least as far as I know), their network of synoptic and climate stations includes questionable sites such as  Heathrow, Faversham, a whole host of urban locations, not to mention Motherwell, which as we know is in the middle of a car park.


What we can say though is that this summer the heat was nowhere near as intense as 1976, or other recent summers:



In 1976, there were nine days in CET that were over 30C, with the highest still a record at 33.2C. This year only one day topped 30C, and that was only 30.7C.

The summers of 1975, 1995, 2003 and 2006 all featured much higher temperatures:

Days > 30
1975 4
1976 9
1995 5
2003 2
2006 3
2018 1


Laughingly, the Met Office refer to the fact that hot summers were not as common pre 20thC. Have they not heard of the Little Ice Age?


With wall to wall sunshine for the best part of three months, this summer should have beaten all records, if global warming had anything to do with it. The fact that it did not even get near to 1826 rather says it all!

  1. MrGrimNasty permalink
    September 1, 2018 6:36 pm

    I noticed that the August 2018 CET dropped from near record to completely ordinary – 67th tied with a lot of other years. This was obviously because the weather changed, but also I see that they have entered the final figure as 16.6C, whereas to the 30th it was 17.1C. Another fairly large final adjustment?

  2. Mack permalink
    September 1, 2018 6:41 pm

    1666, in the midst of the Little Ice Age, would probably also have given this summer a run for its’ money. A quick perusal of historical eye witness evidence across Europe, not least the diary of Samuel Pepys, demonstrates that it was a proper scorcher. No doubt caused by the Co2 emissions of the Stuart version of Mr Whippy ice cream vans, I mean nags, working overtime!

    • Peter F Gill permalink
      September 1, 2018 6:54 pm

      Yes Mack. The drought in London lasted most of 1666 which is probably why we had the disastrous event that started in Pudding Lane.

      • Mack permalink
        September 1, 2018 7:19 pm

        You are right Peter. 1665, the year of the ‘Great Plague’, was also on the very hot side. Those pesky rats and fleas do thrive in very warm conditions. This is probably the point where someone with a more encyclopaedic knowledge of H. Lamb’s work on climate history than mine should jump in and explain why the Met Office’s claims and the word ‘unprecedented’ should never appear in the same sentence.

      • Adrian (under-funded). permalink
        September 2, 2018 6:31 pm

        There you go again Paul, using actual data.

        When will you realise that is so so passé. Modern science is done by those that know, really know, what has, is, and will happen because they are so clever. Data are just awkward numbers to be adjusted, massaged or, as required beaten or deleted until they prove what the clever ones know.

        Get with the agenda Paul, you’ll get more research funding.

  3. Tom Dowter permalink
    September 1, 2018 7:11 pm

    For the purposes of the CET series, The Met Office describes Central England as: “a roughly (!) triangular area of the United Kingdom enclosed by Lancashire (!), London and Bristol”. This is a long way from being the same as Britain/the UK. Even the stations that are used to make up the CET have changed. For example, “London” has variously been Greenwich, Kew and Rothampstead.

    For Britain as a whole, records only really go back to 1910. To be sure, we do have some earlier data. For example, when May was being described as the hottest ever for BRITAIN, there were five stations in the GHCN database which had data for 1833 which was the hottest May in the CET series. One of those outside the CET area, Edinburgh Royal Observatory, also had 1833 as the hottest. One had 1833 tying for first place and the other had it in second place.

    The CET series is interesting, but it is hardly global!

    • September 1, 2018 7:20 pm

      As I say, Scotland , Wales and N Ireland have been relatively cooler this summer in terms of anomalies, so you would expect UK to be cooler than CET relatively.

      One of the things about CET though is that when stations change, they are carefully homogenised. In other words, if Rothamsted is 1C cooler than Kew on average, its actual temp is adjusted up by 1C (or the historical database adj down) so that both stations can be directly compared.

      Does the Met Office do the same with the UK calculation? I somehow doubt it.

      • MrGrimNasty permalink
        September 1, 2018 8:01 pm

        Don’t forget that there is a small UHI adjustment for the CET in latter years.
        Did the MO ever increase it in the 2017 ‘overhaul’ as would appear necessary?

      • Tom Dowter permalink
        September 1, 2018 8:39 pm

        Some years ago, I did a study on the reliability of the homogenisation process of the type that you describe using US data. Using neighbouring stations which had a continuous record of parallel running, I investigated how sensitive the apparent temperature increase/decrease would be if one switched from one station to the other at different times. I found that, at the individual station level, the results were rather poor. However, when I switched several such stations at the same time, the errors tended to cancel each other out. This is more or less what one would expect.

        What we have in the case of the CET versus Britain is low sampling and careful homogenisation compared with much more sampling and possibly less careful homogenisation. Who can say, a priori, which is better?

      • September 1, 2018 10:08 pm

        I have no faith at all in the UK series, because the Met Office do not explain how they get to their figures.

        We know that many of their sites are heavily affected by UHI, but do you know if they make allowance for that?

        There are also sites like Faversham, that has only been around a few years, is in a local hot spots and has been judged as being utterly unsuitable by an expert like the late Philip Eden.

        Of course, if you know how the UK dataset is compiled, you would no doubt be able to justify how it is better than long running, high quality sites.

  4. September 1, 2018 7:20 pm

    A quick glance at the heat map and it looks like the places with the hottest temps were mainly a cluster around greater London and around Manchester. UHI?

  5. Alaskan sea permalink
    September 1, 2018 8:40 pm

    There’s still snow here despite the warm summer. The sphinx snow patch, Braeriach in the Cairngorms is 7m long and 5m wide and 0.6m deep.

  6. Jack Broughton permalink
    September 1, 2018 8:41 pm

    Back in 1976 the big fear campaign regarding CO2 was global cooling: a good example of how CO2 has little role in the modern climate. The theory then was that the air would cool more rapidly due to the increased CO2

    The Radiative Forcing Factors that drove the models then were seriously flawed and still are. They are also very subjective interpretations of atmospheric transmission of radiation: certainly not a good basis for doomsday modelling.

    The fault is the belief that carbon dioxide emits the same when mixed with moisture as it does alone: it does not!

    • richard verney permalink
      September 1, 2018 10:03 pm

      That assumption might be right, and I have never seen a proper estimate of this.

      It is essentially a question of the ratio of the number of incoming photons in the three CO2 absorption bands (of 2.7, 4.3 and 15 µm wavelength), compared to the number of photons in those bands emitted by the surface, but adjusted by the fact that the surface emits 24 hrs per day whereas solar only illuminates one half of the sphere at any given time.

      A CO2 molecule does not know whether say a 15 µm photon is incoming from solar irradiance or upwelling from surface emissions. Thus if there are more incoming 15 µm photons than there are upwelling 15 µm photons, it follows that increasing CO2 will backradiate towards space more of the incoming 15 µm photons, than it will backradiate towards the surface, the upwelling 15 µm photons that the surface had emitted.

      Now although solar irradiance does not contain much in the way of LWIR, this is all relative. It is a facet of Plank’s Law that the hotter the source, the more photons will be generated. See:

      Thus although the 2.7 µm. 4.3 µm and 15 µm photons are in the wings of the EMR spectrum as emitted by the sun, there will nonetheless be more photons of these wavelengths than photons of similar wavelength emitted by the surface, simply because the sun is a hotter and more intense source.

      One can see this when looking at the above plot, You will note that at all times the blue line (5000K approx temp of the sun) is above the green and the red lines. The surface of the Earth emits at an average of about 288K (ie., about 15 deg C), and if this spectrum was to be plotted it would be lower and flatter than the red line (3000K plot) and would peak at 10 µm (which incidentally is not a wavelength band that CO2 absorbs).

      PS. One should also not overlook that the absorption spectra of water vapour and CO2 overlap quite considerably, but that is a different issue

      • Jack Broughton permalink
        September 3, 2018 6:53 pm

        It is the “overlap” that causes much of the error: the calculations assume that the two coefficients are simply additive. The absorption by CO2 and H2O are not simply additive: the emission of a mixture is less than the sum of the two parts as has been known since the work of HC Hottel in the 1940s.

        Another approach to the effect that is often used, is to add the two partial pressures and treat the mixture as a single component gas. The product of path-length and partial pressure greatly reduces the absorption coefficient compared with adding the two single components.

        The RFF as used by IPCC assumes the simple “Grey gas” model is correct for the two components. The whole issue of RFF assessment is a fiasco in the IPCC reports, yet underpins the massive fear campaign because so few people ever question it.

  7. Gerry, England permalink
    September 1, 2018 9:23 pm

    2003 and 2006 matching 1976? Well, I don’t recall them being close to 1976. Maybe it was the ice-cream vans parking by the temp stations.

    • MrGrimNasty permalink
      September 1, 2018 10:00 pm

      Raw stats. don’t really tell the whole story do they. You might have similarly warm summers but one could have very very hot spells interspersed with cold, and another may be consistently warm for exactly 3 months. That’s why obsessing over records and looking for ‘meaning’ is daft. Months and summer are defined by man, change the calendar by a day and all the records change!

  8. September 1, 2018 9:24 pm

    Reblogged this on WeatherAction News and commented:
    It’s quite likely as Climate Realists had stated that UHI is behind much of this “record”. Much of June had relatively cool sub 10°C nights, with the centre of town a good 2-3°C warmer (guess where the official temperatures are recorded?). The DM noted back in June about when London had an uncomfortable night with an 18°C min, whilst the surrounding countryside was cozy under 10°C.1995, 2003 and 2006, which I spent in London, certainly felt hotter than this year and those who lived through 1976 have 2018 not even close. It was a great summer though, yet again caused by blocking.

  9. September 1, 2018 9:26 pm

    A quick look at Iceland’s weather this year gives the lie to the global warming crackpots – that means you too, Met Office.

  10. MrGrimNasty permalink
    September 1, 2018 10:07 pm

    UAH for Aug 2018 dropped back again +0.19C now, been very consistent all year.

    • dave permalink
      September 2, 2018 8:42 am

      “UAH…dropped back…consistent all year.”

      Yes Daily figures for the atmosphere and the SST have been very steady for six months..

      One can look at the following, regularly, and thereafter rarely ‘get a jolt’ from the monthly figures from UAH and RSS:

  11. Tom Dowter permalink
    September 2, 2018 12:51 am

    Replying to Paul, above . . . .

    Well I guess “you pays yer money and you takes yer choice”! Personally, I have little faith in either series – especially when dealing with extremes like the hottest/coldest whatsit evah over relatively short periods like a month, a season or a single year.

    I find that one learns more from examining the broad sweep of things using larger chunks of data.

    If the Met Office were really trying to produce propaganda supporting a warming meme, they would concentrate on the CET. If one takes a ten year moving average over the whole period covered by the series one finds that 18 out of the 20 hottest decades end in this millennium. The two that spoil the full house of 18/18 end in 1998 and 1999!

    Of course, this tells us nothing about what caused the warming and how much of it, if any, is down to increased CO2 in the atmosphere.

    • Mack permalink
      September 2, 2018 2:04 am

      Tom, I’m unaware of any provable data that demonstrates, within a shadow of a doubt, that anthropogenic carbon emissions have made any significant impact to the earth’s climate at all in the last few decades as opposed to what we would normally expect from natural variability caused by solar, lunar and ocean current cycles. Notwithstanding our ascent from the Little Ice Age,
      and all the warmth such a transition naturally engenders, like a true skeptic, I await to see the empirical evidence that will convince me otherwise. Yes, we humans may have added a little warmth to the climate by our activities, mostly through UHI effects, but we still only occupy approx 1% of the planet and Mother Nature holds all the cards. It is only hubris, guilt, or a cynical desire to change our way of life under more sinister and hidden controls, that would convince anyone, with a full set of marbles, that our current climate is not relatively benign and pleasant.

    • Phoenix44 permalink
      September 2, 2018 10:15 am

      I think the Met Office, like the Treasury, retains a professional attitude towards actual data, perhaps not shared by others like BOM. It’s how they then interpret and “explain” the data that I find dubious, and, like the Treasury, their forecasts just reflect their biases.

  12. JCalvertN permalink
    September 2, 2018 2:09 am

    Do they use the Historic Stations set?
    Oxford (1853), Southampton (1855-2000), Armagh (1866), Stornoway (1873), Durham (1880), Sheffield (1883), Bradford (1908), Eskdalemuir (1914), Wick Airport (1914), Lowestoft (1922), Lerwick (1930), Nairn (1931), Ross-On-Wye (1931), Tiree (1931), Valley (1931), Manston (1935), Aberporth (1942), Waddington (1947), Heathrow (1948), Ringway (1956-2004), Hurn (1957), Leuchars (1957), Shawbury (1957), Braemar (1959), CambridgeNIAB (1959), Cardiff Bute Park (1959), Cwmystwyth (1959-2011), Eastbourne (1959), Newton Rigg (1959), Paisley (1959), Sutton Bonington (1959),
    Ballypatrick Forest (1961), Whitby (1962), Yeovilton (1964), Dunstaffnage (1972), Camborne (1979)
    A ridiculous proportion of these are at airports or airforce bases. And Eastbourne is sited in the middle of a city roundabout alongside three electricity substations.

  13. September 2, 2018 5:39 am

    Also seen in the CET are (1) differences in the rate of warming among the calendar months and that (2) the assumption of static and unchanging seasonal cycle in the construction of temperature anomalies is not supported by the data.

    Please see:

    Your comments appreciated

  14. Stonyground permalink
    September 2, 2018 8:29 am

    There was non of this hysteria in 1976 either. It was just a nice hot summer and we just got on with enjoying it.

    • richard verney permalink
      September 2, 2018 11:29 am

      If these summers come around approximately every 40 years, one is lucky to enjoy only 2 such summers in one’s life. The UK could be really nice if it were say 2 to 3 degC warmer when it would be more like Brittany/Normandy temps. The UK would have to warm close to 8 to 10 deg C to get to Spanish temps.

      I am sure that Scotland would love to have the temperatures seen in the Midlands, the Midlands to have the temperatures seen in the South West, the South West to have the temps seen in the Channel Isles and the South East to have the temps seen in Brittany/Normandy. That would be a real boost for the UK tourist industry.

      • mikewaite permalink
        September 2, 2018 7:18 pm

        Just to demonstrate how ridiculous modern journalism has become the saturday daily telegraph a few weeks ago ran 2 articles almost back to back . One of them deplored the hot sunny weather ( the school holidays and the traditional cold wet weather had yet to start) , calling it “hell on earth”. The next page ran articles advising people to invest in homes in “sun drenched Puglia” ( where the summer temperature was about 40C, ie about 10c more than London at its worst (or best?)).
        Of course it could have been a clever idea of the DT editor to show the silliness of global warming madness – but actually I do not think DT editors can cope mentally with that concept.

  15. saparonia permalink
    September 3, 2018 7:21 pm

    There has been hardly any rain at all where I live near Sheffield. The map shows moderate rainfall, we’ve not had any long enough thoroughly wet the ground since the end of June. You only have to look on google maps to see how little there’s been.

    • saparonia permalink
      September 3, 2018 7:23 pm

      Apart from that grumble. I remember Summers as a child, trotting after my Dad and his garden hose.

  16. Jack Broughton permalink
    September 5, 2018 12:05 pm

    I’ve just noticed that table shows how much the sunshine hours have increased, yet no one comments about how that could cause the increased temperatures noted; how odd, especially as there is no known linkage between insolation and CO2, but the UHI is very sunshine dependent. The higher insolation is due to the general adoption of “Clean Air Acts” in the west, (together with lower rainfall of course)

  17. September 5, 2018 5:40 pm

    The BBC covered this extensively, often stating that it was the “hottest summer ever” in England when they meant “hottest since 1910”, no doubt leaving the majority of viewers, who are unaware of the existence of CET, the impression that the former was the case.
    They also quoted the temperature at Heathrow, which is warming faster than CET and is not a reliable measure of the temperature in the rest of the country.

Comments are closed.

%d bloggers like this: