Why UHI cools the past !

The Urban Heat Island(UHI) effect ‘cools’ the past in CRUTEM4 land temperature series. This may seem counter-intuitive but the inclusion of stations in  large cities has introduced a long term bias in normalised  anomalies.

CRUTEM4 calculated with and without the 500 fastest warming bstations (mostly large cities)

CRUTEM4 calculated with (red)  and without(blue) the 500 fastest warming stations (mostly large cities)

The reason for this bias is that each station gets normalised to the same 1961-1990 period independent of its relative temperature. Even though we know that a large city like Milan is on average 3C warmer than the surrounding area, it makes no difference to the apparent anomaly change. That is because all net warming due to city growth effectively gets normalised out when the seasonal average is subtracted. As a direct result such ‘warm’ cities appear to be far ‘cooler’ than the surrounding areas before 1950. This is just another artifact of using anomalies rather than absolute temperatures.

I analysed all 6520 stations in CRUTEM4 and identified those stations whose average anomaly increase from the period before 1921 to the period 1990-2015 was greater than 1C . There are 538 such stations, the majority of which are in urban areas. These include places like San Diego, Calcutta, Melbourne, Beijing, Shanghai etc. The full list is here. I repeated exactly the same calculation of CRUTEM4 annual anomalies both with and without these stations. Even without these cities the remaining coverage is almost unchanged at 5982 stations. The results are shown above and below.

Detailed look at the change in global temperature anomalies before 1950. Excluding Cities increases 19th century anomalies and reduces net global warming.

Detailed look at the change in global temperature anomalies before 1950. Excluding Cities increases 19th century anomalies and reduces net global warming.

The most rapidly changing stations due to post 1950 urbanisation cause a net reduction in 19th century anomalies of ~0.2C

Detail of recent trenbds in land anol,amies with and without the 538 'cities'.

Detail of recent trenbds in land anol,amies with and without the 538 ‘cities’.

The recent trends are very much smaller – only ~0.02C. The reason for this is simply that the fastest ‘energy’ growth in major cities had already occured before 1990. The overall rise in temperature gets normalised out when calculating anomalies. However its effect simply reappears as an excess ‘cooling’ in the 19th and early 20th century.

Past studies (including mine!) have claimed that the UHI effect is very small partly because they focus on  recent trends in temperature anomalies. However this is not the case once  the pivot effect of the normalisation period is included. Overall I find that the UHI has increased global warming on land by about 0.2C since 1850 by artificially supressing land temperature anomalies pre-1960.

This entry was posted in AGW, Climate Change, Institiutions, UK Met Office and tagged , . Bookmark the permalink.

17 Responses to Why UHI cools the past !

  1. Eric Barnes says:

    Nice work Clive! Thanks!

  2. Nick Stokes says:

    “As a direct result such ‘warm’ cities appear to be far ‘cooler’ than the surrounding areas before 1950.”
    This has nothing to do with anomalies or UHI. It simply expresses the fact that they are warming, for whatever reason. Sites at a point in time can be locally warmer or cooler for all sorts of reasons, UHI, altitude, N-S hillside, sea-side vs inland. Anomalies are used to remove this effect, whatever caused it.

    “However this is not the case once the pivot effect of the normalisation period is included. Overall I find that the UHI has increased global warming on land by about 0.2C since 1850 by artificially supressing land temperature anomalies pre-1960.”
    There is no pivot effect. And you haven’t pinned anything on UHI. All you have done is take out the 538 stations that warmed most rapidly, and shown that that makes a difference to warming rate. Well, of course.

    • Clive Best says:

      Nick,

      Have you actually checked the stations listed which have the largest changes in anomaly ? The list includes very many large cities which have grown enormously since 1900 – as you would expect. Just starting on the first page……..

      Tromso, Tronheim,Bergen, Oslo, Uppsala, Stockholm, Helsinki, Lerwick, Aberdeen, Oxford, Durham, Rejkyavik, Copnehagen, Brussels, Luxembourg, Zurich, Geneva, Lugano, Paris, Nancy, Strasbourg, Dijon, Lyon, Toulouse, Montpelier, Nimes, Marseille, Nice, Perpignan, La Coruna, Pamplona, Zaragoza, Barcelona, Salamanca,Madrid ……
      etc.etc.

      I also think you have misunderstood the point. It is not a question as to whether a city is 1C warmer than the surroundung countryside. It is a question as to whether the rate of change in temperature since 1900 has been artificially increased by the rate of change of urbanisation and transport. As you well know that is what anomalies measure.

      A station on a hill can be 3C colder than one in a valley but it is not increasing in height with time. The difference remains constant and the anomalies for both stations will be equal! That is the point.

      Growth in major cities basically occured before 1990. By using a period 1961-1990 for normalising anomalies Cities will not show excess warming post 1990. Instead they will show excess cooling during their previous period of growth.

      I think this is the reason that the UHI effect is not seen in recent data. It only shows up in the past when rapid urbanisation was occuring. All stations renormalise themselves to the fixed period. The terms warming and cooling are allways relative to this.

      Interestingly enough London UK does not appear – perhaps because it was already a teeming city in 1850

  3. Nick Stokes says:

    Clive,
    “The list includes very many large cities which have grown enormously since 1900 – as you would expect.”
    That wasn’t part of the case as you stated it. If you want to establish that, you have to give and analyse the numbers.

    You chose 538 stations, not because of analysed UHI liability, but because they had warmed 1C. Then because the average without them warms less, said that says something about UHI. It doesn’t. If you take the five tallest kids out of a class the average height goes down. That doesn’t mean they were on steroids.

    “By using a period 1961-1990 for normalising anomalies Cities will not show excess warming post 1990.”
    They won’t show a warming relative to 1961-90. But they will show the same rate of warming in that time, independent of what period you chose. And the warming relative to 1961-90 will be the same with any other base.

    But the talk about when cities grew is hand-waving. London grew rapidly early. But the CRUTEM stations are Heathrow and Gatwick. They would have been pretty rural in 1920.

  4. Clive Best says:

    If you take the five tallest kids out of a class the average height goes down. That doesn’t mean they were on steroids.

    No – but if you discover afterwards that they were on steroids then this might be part of the explanation as to why average height in the school is increasing with time.

    I agree this should be done more rigidly but my basic point stands. Any UHI effect is concentrated in the period of most rapid growth especially in cars/lorries. That occured mostly before 1961-1990.

    • Eric Barnes says:

      To answer Nick’s issue, repeat the process (take the next 500 highest trend records and recalculate). If it’s really UHI, you’ll soon run out of UHI stations and get more stability. If Nick is correct, the trend would be somewhat regularly distributed among those stations.

      I guess more generally calculating trend per station and seeing how it is distributed would answer that problem thoroughly.

      Thanks

      Erick

  5. Geoff Sherrington says:

    Admittedly we know less than we would like to know about the quantitative/time effects of UHI. However, it would be imprudent to claim that it has not affected the historic T record.
    It is also imprudent to downplay potential UHI magnitudes by using simple analogies for a complex problem.
    For a slightly complex problem, how is the validity of the anomaly method affected in the hypothetical case that a strong UHI rise took place in the reference period, such as 1961_1990?
    Can anyone refer us to a scenario/model/methodology/correction paper where UHI changes in the reference period?
    Geoff

  6. Ron Graf says:

    Clive, I believe your analysis underscores the difficulty in normalizing anything in climate science.
    It seems predictable that your criticism would be dismissed for being oversimplified.

    Whereas scientist had to define the UHI effect by quantitative analysis of city stations vs. rural, I don’t understand why they would have a problem applying the proper correction. I can imagine the scientists had to correlate population density to the UHIE. If that is the case they could have made a sliding adjustment for UHIE based on that known function.

    Nick Stokes:

    Sites at a point in time can be locally warmer or cooler for all sorts of reasons, UHI, altitude, N-S hillside, sea-side vs inland. Anomalies are used to remove this effect, whatever caused it.

    Are not these effects worthy of being studied and quantified? Instead of having our current modeling resources devoted to making black boxes that have proven to be blind to the future, why not use the vast and growing data to model all the various local effects on temperature? Once we quantified all the effects with high certainty we could evaluate the individual stations for aberration from the expected reporting. We might then find, for example, that the station had an air conditioning condenser installed within 100 feet of the thermometer in 1972. Once that was confirmed the data could be corrected from that specific point forwards for hot days.

    With regards to time of day observation change (TOBS,) if local effects are understood well enough to infill the expected temperature at many of the adjusted stations for readings immediately before and after the changeover one could statistically quantify the effect of TOBS.

  7. pier says:

    the news.. of your partner

    Un saluto cordialissimo a te come a tutto il forum….. 🙂 :cheer: presto dovrei dare delle importanti notizie riguardo la ricerca scientifica all’estero che sta andando benissimo, e sto aspettando un importante esito di peer rewiew eseguito in uno dei più importanti centri mondiali scientifici di divulgazione a livello globale top di ricerche scientifiche innovative…….

    … snip for basic relevance …….

    How do you give credit to such a person
    I have not translated the script because it is’ in a bad Italian …………..

  8. Clive Best says:

    There is indeed a paper under review on this subject which I believe shows strong evidence of an effect. However I really cannot comment any further as it would be inappropriate.

    Not quite sure where your quote is taken from but it looks very much like an internal Italian squabble to me ! Again any discussion is too premature.

    saluti.

    Clive

  9. As Nick said, your approach is something of a tautology. Of course you have an impact on the trend when you select stations based on trend to remove.

    A less biased way would be to use metadata (e.g. something like population density) to remove all stations that exceed a certain threshold.

    We had a poster on this at the AGU awhile back (of which Nick Stokes was one of the authors): http://wattsupwiththat.com/2011/12/05/the-impact-of-urbanization-on-land-temperature-trends/

    We also published a paper in JGR looking at this issue in the U.S. specifically: ftp://ftp.ncdc.noaa.gov/pub/data/ushcn/papers/hausfather-etal2013.pdf

    • Clive Best says:

      I think you and Nick are making the same mistake I did by analysing urban and rural stations separately to derive global anomalies. We already know that cities are warmer than surrounding countryside. All that matters for anomalies is whether they are warming faster simply because anomalies measure DT/dt. So it is only rapidly developing cities that matter over a short time period – not the size of a city. New York grew rapidly in the early 20th century. Sao Paolo and Beijing grew in the last 40 years. This affects anomalies. It is exactly the same effect as moving a station up in altitude. Your (BERKELEY) and NCDC algorithms detect the sudden change in temperature caused by resiting , BUT they don’t detect the slower change due to modern growth (cars,heating etc). So you correct one but not the other.

      Cities developing in the third world from 1950-1990 get nomalised to zero just like they reduce altitude. This then artificially ‘cools’ past anomalies. I estimate that this alone increases global anomalies by ~0.2C prior from 1880 to 1940.

      Anomalies really only measure acceleration. F = ma

    • zeke housewife, there is more to the planet than big cities. Even if not corrupted; one thermometer can tell the temp for 3m2 – not even for a whole room, because in one room is most of the time warmer by a degree close to the sealing, than close to the floor. Therefore: you have correct temperature: ”for 20000m3 on the planet, ONLY for the hottest minute in 24h.- nothing for the rest of 99,999999999999999999% of the planet! IT tells on what you are wasting your life… pity…

    • Zeke Hausfather , for you AND the rest of you honest / honorable boys, I have much more important job: to ”peer review” this post; by most critical eye, but be objective please: https://globalwarmingdenier.wordpress.com/global-warming-lost-its-compass-again/

    • clivebest says:

      Zeke,

      I looked at the poster which was a study of a large number of US stations. You did see a small effect – a warming trend of ~ 0.04C/decade in the US. I think it would be interesting to compare gloabally by including/excluding cities like Mexico City or Beijing which have expanded enormously in recent decades. As you say it is not how large a city is that matters. It is how fast the urban heat is increasing.

Leave a Reply