Urban Heat Island - World tour

Posted by Frank Lansner (frank) on 27th February, 2010
U-V-W >>

Far most temperature measurements over longer periods are located in the USA:

Therefore, possibilities to study UHI (Urban Heat Island) effect on measured temperatures are best when examining the large US temperature data set.

With this challenge in mind we will take a look at UHI effects on temperatures around the world.

(Further introduction to UHI, see)


I fell over this fascinating split up of warming trend in latitude in a writing of Fred Singer.  For the years 1979-95 far most warming in the temperature data happens from 30N to 55N:


Well, in that period, most of the global warming happened over land. So for the warming 1979-1995 the areas that warmed far most in the world are marked with red:

After the 1998 El Nino, a huge pacific warming reached the Arctic and the warming profile in latitude of the world changed shape. But still, the warming profile of 1979-1995 is interesting.


When the Tasman institute in 1991 when through the long temperature series from Australia they split data in just two groups: The biggest towns: "Capital cities" and then all the rest including minor towns and rural stations.

This simple split gave 1 degree Celsius warming trend per century in the capital cities:



AUSTRALIA, Melbourne





There are just a few long stations in Brazil, but let's take a look anyway:

These data does not "prove" much on their own, but obviously, they are in line with the viewpoint, that warming trend is boosted when temperatures are measured from growing cities.

The area of Brazil is comparable with the area of USA. So even though USA has far more temperature stations, the warming trends from a few Brazilian thermometers counts just as much when calculating global temperatures. And indeed it seems that also Brazil has a UHI problem?




UHI results for the Ottawa area, East Canada:




Jones et al concludes approximately 0,53 K UHI pollution in warming trend across China for the period 1951 - 2004 - or "0,1K per decade" in the period.




Europe as a whole, Balling 1997, temperatures corrected for UHI:

A spectacular decline in temperatures after 1940-50.


EUROPE, Holland
One of the highest population concentrations in the world. It seems that the country as a whole is warming faster than the rest of the world:


EUROPE, Poland

The Lodz area shows UHI signal, here data taken from a high quality study: http://www.springerlink.com/content/nkm867502t2j8773/fulltext.pdf

The UHI difference in temperatures often reaches 5 degrees Celsius in the summer time, while in the winter time the UHI difference is reduced to approximately 1 degree Celsius.



Simple comparison for Madrid area showing UHI effect on 2 degrees Celsius over 50 years:


 JAPAN, Osaka

Simple comparison for the Osaka area showing UHI effect on 2 degrees Celsius over 50 years:

JAPAN, Tokyo

Simple comparison for the Tokyo area showing UHI effect on 2,5 degrees Celsius over a century:



Warwick Hughes: "This paper by Dr Robert Balling Jr and myself examines a series of rural and small town South African temperature data and compared 30 year trends with available Jones 1994 grid box trends."

So this is a comparison of Rural sites with ALL sites, which shows how much the overall graph is from the more correct rural temperature trend. The difference over approximate 30 years adds up to 0,7 degrees Celsius:



SOUTH AFRICA, Johannesburg

Simple comparison for the Johannesburg area showing UHI effect on 1 degree Celsius over a century:




6 pairs of urban-rural temperature set where analysed:

Here are the locations:

The UHI difference over just 30 years is approximately 1 degree Celsius, with the strongest warming trend for the towns with the biggest relative growth in the period 1970-2000.


USA, all - Thomas Karl

The largest analysis of UHI in USA and in the world. Many hundred Urban-Rural pairs where analysed with a clean systematic and convincing result for the UHI effect on temperature measurements:


ALL USA, Dr. Edward Long 

- Compares one Raw Rural station and one Urban station for 48 different US states and finds:

Raw Rural temperature data (the only valid data set) after the temperature decline 1935-70 never reaches the warm 1935 temperatures again. Many years 1930-50 are warmer in the USA than any recent year.


USA, all Karl + Jones 1989

Here a graph of overall USA temperatures corrected from UHI. In 1984 the Rural temperature of USA is still 0,5 degrees Celsius lower than 1934:



USA, Arizona:

The Rural Gran Canyon is located rather close to the large city Tucson.

Like most Rural USA temperature, Gran Canyon peaks in temperatures in the 1930´ies. The UHI error for Tucson can be estimated to 2 - 3 degrees Celsius over a century.


USA, California:

- Again, a solid connection between population of an area and the warming error in temperature trends. The unit is degrees Celsius PER DEKADE, so these influences are impressive.


USA, Phoenix

Simple comparison for the Phoenix area showing UHI effect approximately 2-3 degrees Celsius from year 1960 to 2000:




USA, New York:

Simple comparison for the New York area showing UHI effect on 2,5 degrees Celsius from year 1900 to 2000:


 See also:


Last changed: 28th February, 2010 at 04:20:29



None Found

Add Comment
Powered by Website Baker