Skip to content

Is The GISS Adjustment For UHI Adequate?

May 20, 2014

By Paul Homewood


Ronan Connolly has sent me a copy of his recent paper “Urbanization bias II. An assessment of the NASA GISS urbanization adjustment method”.





NASA GISS are currently the only group calculating global temperature estimates that explicitly adjust their weather station data for urbanization biases. In this study, their urbanization adjustment procedure was considered.
A number of serious problems were found with their urbanization adjustments:

1.) The vast majority of their adjustments involved correcting for urban cooling", whereas urbanization bias is predominantly a warming bias.

2.) The net effect of their adjustments on their global temperature estimates was unrealistically low, particularly for recent decades, when urbanization bias is expected to have increased.

3.) When a sample of highly urbanized stations was tested, the adjustments successfully removed warming bias for the 1895-1980 period, but left the 1980s-2000s period effectively unadjusted.

In an attempt to explain these unexpected problems, a critical assessment of their adjustment procedure was carried out. Several serious  flaws in their procedure were identified, and recommendations to overcome these  flaws were given.
Overall, NASA GISS’ urbanization adjustments were found to be seriously  flawed, unreliable and inadequate. Until their adjustment approach is substantially improved, their global temperature estimates should be treated with considerable caution.


The paper is very detailed, but I will try to highlight some of the issues Ronan raises.


1) In essence, the GISS homogenisation adjustment, designed to correct for UHI, compares temperature trends at each “urban” site with at least three neighbouring rural sites (that can be up to 1000km away). The trend at the urban site is adjusted to those of the rural stations.

If no such rural sites exist, the urban station is excluded from their global estimates.

2) Although urban sites only cover about 1% of the Earth’s land surface area, about half of the weather stations used for constructing global temperature estimates are located in or near urban areas.

3) Despite this overwhelming imbalance, the GISS UHI adjustment only reduces global temperature trends by 0.05C/century.

4) The reason for this is that a large number of cooling adjustments offset the warming ones. In other words, the GISS algorithm assumes that many urban sites have undergone a cooling bias.

5) Most of the adjustments, that have been made, apply to pre-1970. Since 1970, despite rapid population growth, net urbanisation adjustments are almost zero.


Some of the specific flaws identified by Ronan include:

1) The use of the “extension rule”.

Although GISS drop urban sites where there are not three rural stations nearby, with at least 20 years of records, if the overlap between the urban record and its rural average is shorter than the urban record, they can include some of the longer part of the urban record effectively unadjusted.

So, it is possible, for instance, where the rural record stops in 1970, the urban site would have no adjustment made since 1970, only before.

2) Their method for identifying stations as urban assumes that the co-ordinates they have for the stations are accurate, but quite a few of their station co-ordinates are inaccurate.

As a result, some urban sites are being classified as rural.

3) The identification of urban/rural sites is based on the “Brightness Index”. While this may be appropriate for the USA, which it was originally developed for, it is clearly not suitable for much of the rest of the world.

The paper offers examples in the Middle East and India of major urban sites, such as Baghdad, which have been classified as rural.

This creates two problems:-

a) These sites are not subject to adjustment, as the system thinks they are rural.

b) When they are linked to other urban sites, to compare trends, the subsequent UHI adjustment is incorrect and almost certainly understated.

4) There is an acute shortage of genuinely rural stations with long term and complete records.

5) Since 2011, GISS have used the GHCN V3, which itself homogenises data. This can lead to the transfer of urban biases to rural sites, through a process of blending.

Use of this homogenised data clearly corrupts any urbanisation adjustment GISS make.


As the authors conclude

Until more adequate attempts have been made to  remove (or at least substantially reduce) the non-climatic biases from NASA GISS’ global temperature estimates, they should be treated with considerable caution.

  1. Joe Public permalink
    May 20, 2014 9:05 pm

    “1.) The vast majority of their adjustments involved correcting for urban cooling”, whereas urbanization bias is predominantly a warming bias.”

    I struggle to understand how an urban environment can create ‘cooling’.

    Putting it bluntly, wherever fuel is burnt or power used (i.e. an ‘urban’ environment), the result is exothermic.

  2. Don permalink
    May 20, 2014 9:17 pm

  3. May 20, 2014 9:41 pm

    Paul forgive the o/t

    In spite of this global surge in extreme weather events, these floods, as well as the severe February snowstorm that affected northern Serbia, have created fertile ground for conspiracy theories, such as blaming environmental hazards in the Balkans to a U.S. military-funded project called the High-frequency Active Auroral Research Program (HAARP), located on remote tundra in Alaska. In all of its absurdity, it gets worse, since the Patriarch of the Serbian Orthodox Church has suggested that God is punishing the Serbian people because of the gay parade that took place in September 2013. Unfortunately, a lot less attention is given to the less ‘entertaining’ but real, irreversible process of climate change that is affecting the world

    I find the irony of seamlessly swapping one absurd conspiracy theory, based upon a perverse ideological belief, directly for another quite rich. I’d laugh but these lunatics get employed and are trying to push policy. Hopefully she’ll stay a fiction writer

    London Writers’ Café

    London, GB
    2,888 Creative writers

    Hello, are you a creative writer/author based in London? And looking for a group of 2,500+ like-minded individuals? Congratulations. You’ve found it.London Writers’ Cafe is a…

    Next Meetup

    Cover Letter Critique LIVE with Literary Agent Kate Burke

    Monday, Sep 1, 2014, 7:00 PM
    10 Attending

    Check out this Meetup Group →

  4. May 20, 2014 9:48 pm

    Reblogged this on CraigM350 and commented:
    I’m reading Lamb’s climate papers from the 50s & 60s and it’s notable how often it is stressed that records must be tempered with observations – although he expresses a degree of faith in then modern observations. Plus ça change.

  5. May 20, 2014 11:34 pm

    Thanks, Paul. Interesting article with more warnings and cautions on the use of thermometer records.

    “5) Most of the adjustments, that have been made, apply to pre-1970. Since 1970, despite rapid population growth, net urbanisation adjustments are almost zero.”
    But this is a not a bug, just a feature. 😉

  6. Sceptical Me permalink
    May 22, 2014 7:43 am

    Several years ago, when I first explored climate ‘science’, there were many good empirical studies of the UHI effect and its correlation with population numbers. One study even deduced and proposed a logarithmic adjustment factor for correcting urban data based upon population. The studies demonstrated that rural temperatures increase with the slightest human intervention with the natural world reaching an asymptote, or fixed upper value, as population increases. The logarithmic increase due to human activity have the greatest effect during the initial stages.
    Clearly climate ‘scientists’ have no reason for implementing adjustments that are not consistent with their political agenda and that would undermine their funding.

Comments are closed.

%d bloggers like this: