Table 5.1 Summary of the characteristics of surface, MSU, and radiosonde observations.

Surface

MSU

Radiosonde

Method of observations

Thermometers in shelters (air) or sea water. Since 1982, satellite infrared (IR) oceanic observations blended with in situ observations.

Atmospheric oxygen emits microwave radiation, the intensity of which is measured by the MSU and is proportional to temperature.

Temperature sensors are carried upward through the atmosphere by the balloons and the data are radio-transmitted to ground receiving stations.

Spatial coverage of measurements

Good in most inhabited regions and shipping lanes. Spares elsewhere.

Virtually complete global coverage. Very broad vertical layers (~ km).

Poor in oceanic regions, in the developing world, and in sparsely populated land areas. Good elsewhere. Good vertical resolution from the surface to the lower stratosphere.

Length of observation record

Beginning in mid-nineteenth century, with expanding coverage in first half of twentieth century. Diminished land stations coverage in 1990s.

Begins December 1978.

Beginning in the mid- 1940s, with greatly expanded coverage in the early 1960s, but suffering some deterioration in the 1990s

Directness of the temperature measurement

Direct, in situ observation of temperature blended with satellite infrared for sea surfaces temperature.

Remote measurement of radiative emissions.

Direct, in situ observations of temperature

Time-varying biases

Raw data are influenced by changes in instruments, observing practices, and land use.

Many biases related to, for example, spacecraft altitude, east-west orbital drift, solar heating, and instrument malfunctions.

Many changes in instrumentation, observing methods, and the global network of stations.

Multiplicity of instruments

Sea surface temperature, marine air temperature, and land air temperature measured by tens of thousands of different thermometers of various types.

Usually two spacecraft in orbit; 30,000 observations per day from each; 9 different satellites from 1979 to 1999.

Each sounding made with a new instrument. Dozens of types used over time, varying from country to country, station to station.

Number of independent efforts to produce the data sets

Several groups, employing different methodologies.

One main effort to date.

A few groups, employing different methodologies.