Geoscience Reference
In-Depth Information
Temperature
Temperatures for the last 2,000 years have been estimated from a number of sources, both
direct thermometer-based and proxy-based indicators. What is a proxy? As used here and
elsewhere, it is short for 'proxy variable'. The term 'proxy' is commonly used to describe a
stand-in or substitute, as in 'proxy vote' or 'fighting by proxy'. In the same way, 'proxy
variable' in the parlance of climatology means a measurable 'descriptor' that stands in for a
desired (but unobservable) variable, such as past ocean or land temperature. So there is an
assumption that you can use the proxy variable to estimate a climatic variable that you can-
not measure directly. For example, infrared satellite measurements are examples of a proxy
that can be used to estimating surface temperatures.
Thermometer-based measurements of air temperature have been recorded at a number of
sites in North America and Europe as far back as 1760. The number of observation sites
did not increase to sufficient worldwide geographical coverage to permit a global land av-
erage to be calculated until about the middle of the 19th century. Sea-surface temperatures
(SSTs) and marine air temperatures (MATs) were systematically recorded by ships from the
mid-19th century, but even today the coverage of the Southern Hemisphere is extremely
poor. All these data sets require various corrections to account for changing conditions and
measurement techniques. For example, for land data each station has been examined to en-
sure that conditions have not varied through time as a result of changes in the measurement
site, instruments used, instrument shelters, or the way monthly averages were computed, or
the growth of cities around the sites, which leads to warmer temperatures caused by the
urban heat island effect. In the IPCC science report, the influence of the urban heat island is
acknowledged as real but negligible for the global temperature compilation (less than
0.006°C).
For SST and MAT, there are a number of corrections that have to be applied. First, up to
1941 most SST temperature measurements were made in seawater hoisted on deck in a
bucket. Since 1941, most measurements have been made at the ships' engine water intakes.
Second, between 1856 and 1910 there was a shift from wooden to canvas buckets, which
changes the amount of cooling caused by evaporation that occurs as the water is being
hoisted on deck. In addition, through this period there was a gradual shift from sailing ships
to steamships, which altered the height of the ship decks and the speed of the ships, both of
which can affect the evaporative cooling of the buckets. The other key correction that has
to be made is for the global distribution of meteorological stations through time, which has
varied greatly since 1870. But by making these corrections it is possible to produce a con-
