Figure 17. Surface record for Turahansk. Siberia

 Figure 18 Surface record for Isfjord Radio, Svalbard


Figure 19. Surface record for Islas Juan Fernandez, Chile


 Figure 20. Surface record for Godthaab and Anmgssalik, Greenland

It will be seen that many surface records in remote parts of the world show no signs of warming for the past century. There is some emphasis here of records from the Arctic, as there are persistent false claims that there is warming in the Arctic region, which are contradicted by all available measurements.

The record from Isfjord Radio, Svalbard (Figure 20) is of particular interest, as it illustrates some of the problems in making use of surface records. This record is used by all three surface compilations (GHCN, CRU and GISS), and it is the basis for the large red dot in Figure 1, indicating a temperature rise of 4.1°C between 1901 and 1996. The actual record, which starts in 1910, shows a rise in annual mean temperature from -11.5°C to -4.2°C between 1917 and 1923, an increase of 7.3°C in only 6 years. Yet we are complaining of a mere rise of 0.8°C in 160 years! The inhabitants of Svalbard must have been devastated.. Extreme fluctuations in annual mean recorded temperatures are to be found in many Russian/Siberian records, both in the surface and the MSU records.


The evidence is overwhelming that temperature records from places remote from human habitation show no evidence of warming. These remote places include forests, ice cores and other proxy measurements, measurements by weather balloons, measurements by satellite (the only truly global measurements) and surface measurements in places where human influence is minimal. It only remains, therefore to characterise the human influence around weather stations in more detail;.

Figure 21 shows the equipment currently used for the measurement of surface temperature in many weather stations world wide. It shows the Stevenson screen, a device invented by Robert Stevenson, the Scottish lighthouse engineer and father of the author Robert Louis, in the early 1800s, and has been little changed since. It consists of a wooden box with louvred sides and ventilated roof , painted white, with a front lowering door.

 Figure 21 Surface temperature measuring equipment in current use at the Isle of Man airport, in a Stevenson screen

Temperatures are measured, in this case, by mercury-in-glass thermometers.. Heat transfer to the thermometer takes place by the conventional three mechanisms, conduction, convection and radiation. The conductivity of air is low, so transfer by conduction will be small provided there is air circulation within the screen. However, there will be calm days when the air temperature within the screen, rather than that outside, has an important influence. With good ventilation, convection exchanges heat with the thermometers.

The air entering the box, will however, not necessarily be representative of the outside background climate. If it comes over mountains, open fields or the ocean it might approximate to that background. If it comes from a large urban area it will have exchanged heat with the buildings and other heated surfaces. At an airport it will have been in contact with large aircraft. The effect of “heat islands” and “urbanisation” on weather station readings has long been recognised and many studies have attempted to account for it by comparing measurements in “urban” and “rural” stations, defined rather crudely by differences in population.

The three sets of amalgamated surface readings from the University of East Anglia (UEA), Goddard Institute of Space Studies (GISS) and Global Historical Climate Network (GHCN) all make corrections for “urbanisation” based on comparing “urban” and “rural” sites within a grid box, and reducing the “urban” increase to correspond to the “rural” behaviour. The actual corrections are very small in each case. The system only works for boxes with many weather stations. Areas with only a few boxes in a square, or recent figures where trends are not clear, cannot be corrected. Also there is no recognition of the fact that even so-called “rural” sites, defined, sometimes, as below 10,000 population, are also subject to local convective heating of incoming air. Airports, usually classified as “rural” are often heat islands.

The air entering the Stevenson screen will have a component that has exchanged heat with local building surfaces, ground surfaces, roads, vehicles and aircraft. The example in Figure 22 appears to have an open field behind it, but the circulating air will undergo a sudden temperature rise every time an aircraft is near.

Besides convection, the thermometer exchanges heat by radiation. The caption to this picture from the Internet stated that the white paint on the screen prevents heating effects from the sun’s radiation. This is not true. White paint absorbs between 30 and 50% of the sun’s radiation. The screen will become hot in the sun, and on a calm day, without the ventilation in the roof space, it will dominate the temperature inside. Any deterioration in the paint, such as a loss of gloss or accumulation of dust, will increase the temperature of the screen in the daytime without influencing radiation emission at night.

White paint emits 95% of infra red radiation, so the temperature of the box will be conveyed to the thermometer from the internal surfaces. On a calm day this will be the main influence. The sun’s radiation will also heat local buildings and roads, particularly those that are dark in colour, and this heat will supply radiated heat to the screen and continue to do so, and affect convected air, when the sun has gone. All surfaces will cool by radiation overnight, again depending on their infra red emissivity, which for most surfaces, including those painted white, is 95%

Changes in the heating of neighbouring buildings, darkening of their surfaces, or of ground surfaces (such as by sealing of roads) and increases in number and size of vehicles and aircraft, will all contribute to an upwards temperature bias. An increase in shelter, such as by growth of neighbouring trees, will increase the influence of local convective interchange and of radiation. The composite atmosphere in the screen is a characteristic of the local microclimate, suitable for local weather records and prediction. But it does not measure the local background climate and if the measurements are used to judge changes of surface temperature over many years, then any changes in the surrounding environment will alter (and usually increase) their influence on the record.

In Figure 22 measurements are made by opening the box. This causes a sudden change in air circulation and radiation and thus a drop in temperature. Many weather stations have recently installed automatic measuring equipment which does not involve opening the box for the measurement. This will mean another upwards bias in the record.

Weather stations do not measure the temperature of the background local climate. They measure a mixture of background climate and thermal properties of the local environment. It is the changes in these local thermal properties over long periods of time that are responsible for the observed “global warming”

An indication of the relative importance of these sources of upwards bias may be obtained from Figure 22, which shows the regional temperature increase from 1976 to 1999 during the winter months, December, January and February which showed most of the temperature rises during this period

It will be seen that most of the temperature rises over this period took place in the USA, Europe and Russia, all places with cold winters. The implication, surely, is that the rise resulted predominantly from better heating in the buildings surrounding the weather equipment during a period of improved living standards. Note that there were falls in temperature in the Arctic and Antarctic.

Additional information on the many reasons for upwards bias in most surface measurements is given by Gray (2000) and Daly (2000).


Global temperature measurements remote from human habitation and activity show no evidence of a warming during the last century. Such sites include “proxy” measurements such as tree rings, marine sediments and ice cores, weather balloons and satellite measurements in the lower atmosphere, and many surface sites where human influence is minimal. The small average and highly irregular individual warming displayed by surface measurements is therefore caused by changes in the thermal environment of individual measurement stations over long periods of time, and not by changes in the background climate.

 Figure 22 Temperature rises in 5°x5° latitude/longitude grids from 1976 to 1999 in the winter months December January and February (IPCC 2000 Draft)


I wish acknowledge my debt to John L. Daly whose website is the source of Figures 10 to 16 and 18 to 21


Angell, J.K. 1999 Comparisons of surface and tropospheric temperature trends. Geophys Res Letters:26 (17) 2761-2764

Briffa, KR, PD Jones, FH Schweingruber, TJ Osborn. 1998 Influence of volcanic eruptions on Northern Hemisphere summer temperature over the past 600 years. Nature 393:350-354.

Daly JL (2000) The Surface Record: Global Mean Temperature and how it is determined at sea level

Folland, CK, DE Parker 1995 Quarterly Journal Royal Meteorological Society 121:319-367.

Gray VR (2000) The Surface Temperature Record.

Hansen J and S Lebedeff 1987 Global Trends of measured surface temperature J Geophys Res 92:13345-13372

IPCC 1990 Climate Change (J.T. Houghton. G.J. Jenkins, J.J. Ephraums, Eds ) Cambridge University Press

IPCC 1996 Climate Change 1995 (J.T. Houghton, L.G. Meira Filho, B.A. Callander, N.Harris, A Kattenberg, K.Maskell) Cambridge University Press

Karl, T.R. 1998 Annexe A “Regional Trends and Variations of Temperature and Precipitation” in The Regional Impacts of Climate Change. IPCC WGII, Cambridge University Press

Mann, ME, RS Bradley 1999 Northern Hemisphere Temperatures During the Past Millennium: Inferences, Uncertainties and Limitations. Geophys Res Letters :26 (6) 759-762

Naurzbaev, MM, EA Vaganov 2000. Extraordinary Warmth in the 20th Century? J. Geophys Res 105:7315-7326

Parker, DE, M Gordon, DPN Cullum, DMH Sexton, CK Folland, N Rayner, 1997 A new gridded radiosonde temperature data base and recent temperature trends Geophys Res Letters :24 (12) 1499-1502

Peterson, TC, RS Vose, 1997, An Overview of the Global Historical Climatology Network Temperature Database Bull Amer Meteor Soc 78:2837-2849

Return to `Climate Change Guest Papers' page

Return to "Still Waiting For Greenhouse" main page