126
Chapter 7: The Simulation of Weather Types in GCMs
If IZI < F
If IZI > 2F
If F < IZI < 2F
the flow is straight (i.e. a pure directional type).
the flow is either strongly cyclonic (where IZI > 0)
or strongly anticyclonic (IZI < 0):
the Lamb C or A types.
the flow is partly (anti) cyclonic and partly
directional: a Lamb hybrid type e.g. AS.
If F < 6 and IZI < 6 the flow is light and indeterminate:
a Lamb "unclassified" type, U.
Using this "objective" scheme, each day between the start of December,
1880 to the end of 1989 has been typed (Jones et al., 1993). These daily types
have then been totalled to give monthly indices of the 7 basic types defined
by Lamb, i.e. anticyclonic (A), cyclonic (C), northerly (N), easterly (E),
southerly (S), westerly (W) and northwesterly (NW). The final (NW) type
is calculated separately to conform with Lamb's opinion that NW air masses
are sufficiently distinct to be considered as a major type. In calculating these
major types, hybrid weather types are counted in proportion to the major
types of which they are composed, so that CNW would count half to the C
total and half to the NW total. Together with the "unclassified" types the
resulting totals for each of the 8 types then sum to the number of days in the
month.
Seasonal totals ofLamb's manually classified days and the "objective" days
have been compared in detail for each of the 8 types over the whole period
between 1881 and 1989. They are found to be highly correlated. Individual
scatter plots and comparison time series are shown in Jones et al. (1993). The
correlations are particularly good for synoptic (i.e. A and C) types but they
are not as high for some directional types. This is most likely attributable to
Lamb 's primary concern with the general steering of weather over a succession
of days and his use of upper air charts (only after 1945). Lamb also used
two surface charts per day whereas the objective scheme indices are based on
only one.
The slight differences in the manual and objective indices are not significant
as regards GCM-model validation, however, as the same objective approach
is applied to the classification of the observed and generated pressure fields.
Also, the correlations between the weather types and observed precipitation
and temperature over England and Wales are extremely similar regardless of
whether they are manually or objectively defined (Table 7.1). The following
description of the comparisons between "observed" and model-based weather
type indices are therefore based on the objective classification scheme.
7.4 Details of the Selected GCM Experiments
The comparisons described here are based on the control-run outputs of two
GCMs. A summary ofthe details ofthese models is given in Table 7.2. UKHI
Chapter 7: The Simulation of Weather Types in GCMs
If IZI < F
If IZI > 2F
If F < IZI < 2F
the flow is straight (i.e. a pure directional type).
the flow is either strongly cyclonic (where IZI > 0)
or strongly anticyclonic (IZI < 0):
the Lamb C or A types.
the flow is partly (anti) cyclonic and partly
directional: a Lamb hybrid type e.g. AS.
If F < 6 and IZI < 6 the flow is light and indeterminate:
a Lamb "unclassified" type, U.
Using this "objective" scheme, each day between the start of December,
1880 to the end of 1989 has been typed (Jones et al., 1993). These daily types
have then been totalled to give monthly indices of the 7 basic types defined
by Lamb, i.e. anticyclonic (A), cyclonic (C), northerly (N), easterly (E),
southerly (S), westerly (W) and northwesterly (NW). The final (NW) type
is calculated separately to conform with Lamb's opinion that NW air masses
are sufficiently distinct to be considered as a major type. In calculating these
major types, hybrid weather types are counted in proportion to the major
types of which they are composed, so that CNW would count half to the C
total and half to the NW total. Together with the "unclassified" types the
resulting totals for each of the 8 types then sum to the number of days in the
month.
Seasonal totals ofLamb's manually classified days and the "objective" days
have been compared in detail for each of the 8 types over the whole period
between 1881 and 1989. They are found to be highly correlated. Individual
scatter plots and comparison time series are shown in Jones et al. (1993). The
correlations are particularly good for synoptic (i.e. A and C) types but they
are not as high for some directional types. This is most likely attributable to
Lamb 's primary concern with the general steering of weather over a succession
of days and his use of upper air charts (only after 1945). Lamb also used
two surface charts per day whereas the objective scheme indices are based on
only one.
The slight differences in the manual and objective indices are not significant
as regards GCM-model validation, however, as the same objective approach
is applied to the classification of the observed and generated pressure fields.
Also, the correlations between the weather types and observed precipitation
and temperature over England and Wales are extremely similar regardless of
whether they are manually or objectively defined (Table 7.1). The following
description of the comparisons between "observed" and model-based weather
type indices are therefore based on the objective classification scheme.
7.4 Details of the Selected GCM Experiments
The comparisons described here are based on the control-run outputs of two
GCMs. A summary ofthe details ofthese models is given in Table 7.2. UKHI