56
Chapter 4: The Instrumental Data Record
few countries has the methodology remained consistent since the nineteenth
century (Bradley et al. 1985).
In the routine exchange of data, countries exchange monthly mean temperatures, using whatever practice is applied in the country. Proposals currently
before WMO seek to change the number of meteorological variables routinely
exchanged. When this takes pi ace maximum and minimum temperatures will
be exchanged and it is likely that these will become the preferred method of
calculating mean temperature. Maximum and minimum temperatures will
be extremely useful when considering both the causes and impacts of climate
changes. Using maximum and minimum temperature, however, to calculate
monthly means is extremely susceptible to the time the observation is made
(Baker, 1975; Blackburn, 1983). In the United States, Karl et al. (1986)
have developed a model to correct observations made in the morning and
afternoon to the 24 hour day ending at midnight. The corrections to the
monthly means are called time-o/-observation biases.
4.2.4
Changes in the Station Environment
The growth of towns and cities over the last 200 years around observing
sites can seriously impair the temperature record. Affected sites will appear
warmer than adjacent rural sites. The problem is serious at individual sites
but appears relatively small in averages calculated for large continental-scale
areas (Jones et al. 1990). A correction procedure has been developed for the
contiguous United States by Kar! et al. (1988) and depends on population
size. Population is generally used as a ready surrogate for urbanization but
other indices such as paved area (Oke, 1974), although more useful, are only
available for specific cases. The individual nature of the effect me ans that the
corrections applied to U .S. data by Kar! et al. (1988) will only be reliable
on a regional rather than a site basis. The corrections are also specific to
North America and would not apply to Europe, for example, where urban
development has taken place over a much longer period of time. The problem
is likely to become important in the future in the developing countries as cities
expand rapidly (Oke, 1986).
4.2.5
Precipitation and Pressure Homogeneity
Both of these variables are also affected by severe homogeneity problems. For
precipitation, there are problems assodated with the gauge size, shielding,
height above the ground and the growth of vegetation near the gauge. All
can impair the performance of the gauge, affecting the effidency of the catch.
The biggest problem concerns the measurement of snowfall, where continual
attempts to improve gauge performance affects the long-term homogeneity
of the station records. Rodda (1969) has reviewed the possible errors and
inhomogeneities that can result in precipitation records.
Chapter 4: The Instrumental Data Record
few countries has the methodology remained consistent since the nineteenth
century (Bradley et al. 1985).
In the routine exchange of data, countries exchange monthly mean temperatures, using whatever practice is applied in the country. Proposals currently
before WMO seek to change the number of meteorological variables routinely
exchanged. When this takes pi ace maximum and minimum temperatures will
be exchanged and it is likely that these will become the preferred method of
calculating mean temperature. Maximum and minimum temperatures will
be extremely useful when considering both the causes and impacts of climate
changes. Using maximum and minimum temperature, however, to calculate
monthly means is extremely susceptible to the time the observation is made
(Baker, 1975; Blackburn, 1983). In the United States, Karl et al. (1986)
have developed a model to correct observations made in the morning and
afternoon to the 24 hour day ending at midnight. The corrections to the
monthly means are called time-o/-observation biases.
4.2.4
Changes in the Station Environment
The growth of towns and cities over the last 200 years around observing
sites can seriously impair the temperature record. Affected sites will appear
warmer than adjacent rural sites. The problem is serious at individual sites
but appears relatively small in averages calculated for large continental-scale
areas (Jones et al. 1990). A correction procedure has been developed for the
contiguous United States by Kar! et al. (1988) and depends on population
size. Population is generally used as a ready surrogate for urbanization but
other indices such as paved area (Oke, 1974), although more useful, are only
available for specific cases. The individual nature of the effect me ans that the
corrections applied to U .S. data by Kar! et al. (1988) will only be reliable
on a regional rather than a site basis. The corrections are also specific to
North America and would not apply to Europe, for example, where urban
development has taken place over a much longer period of time. The problem
is likely to become important in the future in the developing countries as cities
expand rapidly (Oke, 1986).
4.2.5
Precipitation and Pressure Homogeneity
Both of these variables are also affected by severe homogeneity problems. For
precipitation, there are problems assodated with the gauge size, shielding,
height above the ground and the growth of vegetation near the gauge. All
can impair the performance of the gauge, affecting the effidency of the catch.
The biggest problem concerns the measurement of snowfall, where continual
attempts to improve gauge performance affects the long-term homogeneity
of the station records. Rodda (1969) has reviewed the possible errors and
inhomogeneities that can result in precipitation records.