EXTENDED-RANGE FORECASTING BY WEATHER TYPES 
By ROBERT D. ELLIOTT 
North American Weather Consultants 
INTRODUCTION 
During the last two decades there has been a con- 
siderable development in the field of extended-range 
forecasting, that is, forecasts for up to a week in ad- 
vance. While it is known that a modest success in fore- 
casting general weather trends for as much as seven 
days in advance has been achieved by several groups 
using techniques of their own development, no 
particular method has gained the widespread acceptance 
in meteorological circles enjoyed by the frontal and air- 
mass analysis techniques of short-range forecasting. 
However, most of the extended-range forecasting 
methods have a number of points in common and these 
will be discussed briefly. 
Meteorologists engaged in effective extended-range 
forecasting must become accustomed to dealing with 
certain forms of representation which summarize the 
circulation characteristics of a given region over a finite 
period of time. This may take the form of fixed-time- 
interval mean charts such as the overlapping five-day 
mean maps used by the Extended Forecast Section of 
the U. S. Weather Bureau [20], or more variable and 
‘natural’ period representations such as composite 
maps [21], weather types [3], or steering patterns [26]. 
In each case the representation involves, basically, the 
concept of mean circulation patterns at the surface 
and/or aloft, covering a fixed or variable time interval 
of several days or more. Even a pure analogue selection 
technique must involve considerations of this nature 
in order to be successful. One-map “thumbprint” selec- 
tion schemes have not proved successful in extended- 
period forecasting, and exhaustive statistical tests by 
Darling [6] show that persistence is the principal factor 
taken into account by this method. 
Ordinarily, the time interval is chosen in such a 
manner that the local circulation patterns of simple 
migratory and transitory cyclones and anticyclones are 
smoothed out, thus delineating the more basic quasi- 
stationary and broad-scale features of the circulation 
pattern which are responsible for the longer-period 
weather characteristics in a given region. The mean 
patterns thus revealed are characterized by large-scale 
cyclones and anticyclones at the surface and by an 
undulatory pattern aloft having longer wave lengths 
than those observed on daily upper-level charts. The 
movements and changes in form of these patterns are 
on the whole slower than those of the individual migra- 
tory systems making up the mean, which at least partly 
compensates for the difficulty of having to forecast so 
far im advance. In this country, methods for forecasting 
these movements and changes in form are based upon 
statistical weather-type prediction schemes [8], the ap- 
plication of the conservation of vorticity principle [20], 
and a number of other techniques which will be con- 
sidered later. 
Once the extended-period mean or type pattern is 
forecast, the anomalies! for the weather elements for 
the time interval in question can be predicted. Con- 
siderable statistical work has been done relating tem- 
perature and precipitation anomalies for an extended 
period with the weather types [3] and with the mean 
isobaric pattern at upper levels [17, 20]. In general, 
extended-period surface temperature anomalies are 
negative behind mean troughs aloft and positive ahead, 
with precipitation and cloudiness at a maximum Just 
ahead of the mean trough. 
The problem of preparing daily prognostic charts 
for up to seven days in advance is considerably more 
difficult. It is of course obvious that the sum of the daily 
prognostic charts should correspond to the extended- 
period prognostic representation; however, there are 
a great variety of ways in which this pattern may be 
made up. A prognostic extended-period upper-level 
chart may be used as a “steering” pattern to fix in a 
general way the paths of migratory cyclones and anti- 
cyclones, the individual systems beimg extrapolated ' 
forward from forecast day with due consideration bemg 
given to the predicted strength of the flow aloft. In 
one technique this is supplemented by various prog- 
nostic zonal and meridional indices [20]. Another tech- 
nique [4] resorts to the use of analogues selected from 
the archives of weather charts in which the mean flow 
pattern matches the predicted pattern. Still another 
method is to use ‘‘ideal” weather types [3] giving the 
characteristic daily surface charts for each day of the 
weather types as determined through statistical study 
of past cases. 
The foregoing paragraphs very briefly outline the 
features which most extended-range forecasting tech- 
niques have in common. These may be summarized as 
follows: 
1. A method for concise representation of circulation 
characteristics during an extended period. 
2. A method for forecasting movement and changes 
in form of circulation features delineated in 1, thus 
permitting the preparation of a prognostic representa- 
tion. 
3. A method for forecasting extended-period anoma- 
lies of the weather elements from information provided 
by 2. 
4. A method for preparing daily prognostic charts 
and forecasts for daily weather elements from 2. 
1. The term ‘‘anomaly,’”’ as used in this article, refers to 
departure from a long-term average over time and has nothing 
to do with the departures from means computed for latitude 
circles. 
834 
