Climatic Data, Instrumental Records

DETECTION OF CURRENT global temperature and reconstruction of past trends rely on data from four sources: surface temperature from weather or climate stations, weather balloons, satellite mounted remote sensors, and proxy sources. Data from the first three categories are known as the instrumental record.

Until recently, measurements of global air-temperature change were based entirely on measurements taken on the ground at weather stations. Three authorities that have taken responsibility for the combined surface record are the Climate Research Unit (CRU) of the University of East Anglia (UEA), the National Aeronautics and Space Administrations (NASA's) Goddard Institute for Space Studies (GISS), and the Global Historical Climate Network (GHCN) run by the U.S. National Océanographie and Atmospheric Administration (NOAA).

Given that the data come from weather stations unevenly distributed over the Earth's surface, mostly in the Northern Hemisphere, mainly on land, close to towns and cities, the question arises as to the extent to which these temperatures can be taken as representative of the atmosphere as a whole.

Modification of the surface by human activity can have a significant effect on local climate near the ground. The best-documented example is the urban heat island effect, in which localized warming due to asphalt and concrete replacing grass and trees can influence data from urban stations. This can make an urban area several degrees C warmer than its rural surroundings.

Several researchers have demonstrated that only very small changes in population are enough to induce a statistically significant local warming effect. Many weather stations are located at airports, which originally were located in rural areas on the outskirts of urban areas. But with rapid population growth, airports have made a transition from rural to heat island and it is difficult for the national agencies responsible for archiving climate data to take these influences into account. The problem has been made worse by the fact that two-thirds of the weather stations operating around 1975, mainly rural, have been closed down. For example, the GHCN network showed a peak of about 15,000 stations around 1970, which declined to 5,000 as stations deemed non-essential for operational purposes were closed down. The loss was mainly of rural stations.

Over 70 percent of the Earths surface is covered by water. Because of this, land-based surface temperature data are supplemented by measurements taken at sea, usually by ships. Scientists have assumed that there is a link between the surface layer temperature of seawater and that of the air above it. Sea surface temperatures are estimated from the temperature of seawater as it is taken aboard as an engine coolant. Sometimes, a bucket tethered to a rope thrown overboard serves the purpose. The bucket is then hauled [Page 237]aboard, and the water temperature is measured with a thermometer. Other sea surface data are taken by combining data from several rather different sources, such as buoys, satellite infrared data, nighttime marine air temperatures measured aboard ships, and measurements made at small island stations.

Each method has problems that result in errors or differences among the data sets. The mean global surface temperature of the Earth shows a warming of 0.3–0.7 degrees C over the past century; this is an increase statistically of about 0.003 to 0.007 degrees C per year, a rate not far from the standard error for the data. The data sets from the various sources (CRU, GISS and GHCN) show slightly dissimilar trends as the data are processed in different ways.

...