122
Chapter 7: The Simulation oE Weather Types in GCMs
tic evidence on the performance of the models which should help in their
development and continued improvement. However, this information is also
important for gauging the degree of confidence that should be attributed to
the regional details of climate impact studies based on the output of particular GCMs.
Several recent papers have addressed different aspects of GCM (controlrun) validation (for examples, see the papers cited in Hulme et al. , 1993).
This Chapter describes one particular approach: that of comparing the details of GCM-simulated daily circulation patterns with those observed during
arecent run of years. The following description is largely a synthesis of results published in two recent papers (Jones et al., 1993; Hulme et al., 1993).
Concentrating on characteristic circulation patterns (or weather types) is a
convenient means of assessing the realism of the simulated atmospheric and
land surface processes. The examples described focus on the region of the
British Isles, so enabling us to make use of an existing body of knowledge
about the classification of daily circulation patterns over the region and their
relationship with primary climate parameters such as temperature and precipitation: primarily based on the synoptic classification scheme devised by
Hubert Lamb (Larnb, 1950; 1972) and the studies which have employed it
(e.g. Murray and Lewis, 1966; Jones and Kelly, 1982; Briffa et al. , 1990.)
First, the Lamb daily classification (the Lamb weather catalogue) is described and then an "objective" scheme for automatically producing Lamb's
weather indices from gridded surface pressure data is outlined and discussed.
After abrief description of the GCMs used in this exercise, model-generated
temperature and precipitation statistics, weather-type frequencies and the relationships between these are compared with those observed in "real-world"
data. A very brief comparison is also made between observed and modelgenerated average spell-lengths for two major weather types and a similar
comparison of modelled versus real-world storm frequencies is illustrated.
7.2 The Lamb Catalogue
The circulation patterns dominating the British Isles (approximately the area
within 50° to 60° N and 100W to 2° E) on each day since 1861 (except in the
1860s when no Sunday maps were produced) have been classified according to
a scheme devised by Hubert Lamb (Lamb 1950, 1972; and updated regularly
by hirn in Climate Monitor). The scheme consists of 26 "weather types"
(plus 1 "unclassified") made up from 9 anticyclonic (A) types: A, ANE, AE,
ASE, AS, ASW, AW, ANW and AN; 9 equivalent cyclonic (C) types: C,
CNE etc.; and the 8 directional types: NE, E, SE, S, SW, W, NW and N.
Figure 7.1, reproduced from Lamb (1972), shows typical examples of different
weather types defined according to Lamb's manual classification of mean-sealevel pressure and (after World War II) upper air charts. The full digitised
classification is available from the Climatic Research Unit. Many studies have
Chapter 7: The Simulation oE Weather Types in GCMs
tic evidence on the performance of the models which should help in their
development and continued improvement. However, this information is also
important for gauging the degree of confidence that should be attributed to
the regional details of climate impact studies based on the output of particular GCMs.
Several recent papers have addressed different aspects of GCM (controlrun) validation (for examples, see the papers cited in Hulme et al. , 1993).
This Chapter describes one particular approach: that of comparing the details of GCM-simulated daily circulation patterns with those observed during
arecent run of years. The following description is largely a synthesis of results published in two recent papers (Jones et al., 1993; Hulme et al., 1993).
Concentrating on characteristic circulation patterns (or weather types) is a
convenient means of assessing the realism of the simulated atmospheric and
land surface processes. The examples described focus on the region of the
British Isles, so enabling us to make use of an existing body of knowledge
about the classification of daily circulation patterns over the region and their
relationship with primary climate parameters such as temperature and precipitation: primarily based on the synoptic classification scheme devised by
Hubert Lamb (Larnb, 1950; 1972) and the studies which have employed it
(e.g. Murray and Lewis, 1966; Jones and Kelly, 1982; Briffa et al. , 1990.)
First, the Lamb daily classification (the Lamb weather catalogue) is described and then an "objective" scheme for automatically producing Lamb's
weather indices from gridded surface pressure data is outlined and discussed.
After abrief description of the GCMs used in this exercise, model-generated
temperature and precipitation statistics, weather-type frequencies and the relationships between these are compared with those observed in "real-world"
data. A very brief comparison is also made between observed and modelgenerated average spell-lengths for two major weather types and a similar
comparison of modelled versus real-world storm frequencies is illustrated.
7.2 The Lamb Catalogue
The circulation patterns dominating the British Isles (approximately the area
within 50° to 60° N and 100W to 2° E) on each day since 1861 (except in the
1860s when no Sunday maps were produced) have been classified according to
a scheme devised by Hubert Lamb (Lamb 1950, 1972; and updated regularly
by hirn in Climate Monitor). The scheme consists of 26 "weather types"
(plus 1 "unclassified") made up from 9 anticyclonic (A) types: A, ANE, AE,
ASE, AS, ASW, AW, ANW and AN; 9 equivalent cyclonic (C) types: C,
CNE etc.; and the 8 directional types: NE, E, SE, S, SW, W, NW and N.
Figure 7.1, reproduced from Lamb (1972), shows typical examples of different
weather types defined according to Lamb's manual classification of mean-sealevel pressure and (after World War II) upper air charts. The full digitised
classification is available from the Climatic Research Unit. Many studies have