Not the
least important function of the unit was the training of weather officers
from the allied military services in the techniques of extended forecasting.
Ten weather officers of the Army Air Forces were given a six-week
course of intensive training in 1944 and subsequently assigned to
active duty in various areas. Instructors from the AAF Weather School
at Chanute Field visited the Section from time to time for special
instruction, as did also meteorologists from the Navy and from the
Canadian Meteorological Office. A weather officer from the British
Admiralty was assigned to the Section for an eight-week period for
special instruction. Training was provided also in techniques of Northern
Hemisphere map analysis and upper-air analysis for officers of the
Army Weather Central which was located for a time in the Weather Bureau
building.
Members
of the staff of the Extended Forecast Section were called upon frequently
for special consultation on extended and long-range forecast problems.
Mr. Namias was assigned to the Army Air Forces Weather Division for
a month asa consultant to work with a meteorological mission from
Russia on problems of long-range forecasting. He served also as chairman
of an AAF committee on Far Eastern Weather Planning.
A list
of titles of major research projects carried on by the Extended Forecast
Section during the war is filed in the Archives.
2.
The "Extrapolation" Method of Forecasting.-In contrast to
the method of forecasting under development by the Extended Forecast
Section, which proceeds by consideration of trends and other relationships
shown by five-day mean maps, is the method developed by the Special
Forecast Section under the guidance of Mr. C. L. Mitchell. The basic
difference in the approach to the forecasting problem lies in the
use by the Special Forecast Section of weather elements on daily maps
and their interrelationship to forecast subsequent weather changes,
rather than mean maps as used by the Extended Forecast Section. These
two Sections were operated side by side early in the war in order
to obtain information on the best features of each method and to provide
an interchange of ideas out of which it was hoped would come a combined
unit having higher skill than could be attained by either unit working
alone. To a considerable extent the interchange of ideas was worthwhile,
and some improvement was noted in the accuracy of both methods. Later
on the tremendous demand for extended forecast service to the Army
and Navy made it necessary to assign the Special Forecast Section
almost exclusively to serving the needs of the Navy for forecasts
and research.
The specialized
needs of the Navy were met by placing the Special Forecast Section
in rooms adjacent to the navy Long-Range Forecast Office in the Weather
Bureau building. This had the further advantage of making it possible
for the Weather Bureau Extended Forecast units to see and use observations
data from areas under military control which for security reasons
could not be given general distribution.
Mr. Mitchell
of this Section also received a citation from Navy Secretary Knox
for his work in forecasting for the invasion of North Africa.
Research
problems arising in the course of the work were handled by a unit
financed largely by funds transferred from the navy. A list of the
major projects is contained in the Archives.
3.
Forecasting From World-Wide Weather Relationships.-One
of the most extensive studies of world-wide atmospheric relations
was made in the period of 1906 to 1936 by Sir Gilbert Walker. This
work consisted in large part of a search for relationships between
the weather of widely different areas of the earth, with the hope
that a sufficiently large time lag would be found between the nature
of the atmospheric circulation in one area and the subsequent weather
in another area that useful forecasts could be made for months and
seasons in advance. The work had received considerable attention,
particularly the concept of a "Southern Oscillation" which represented
an interrelation between seasonal pressures and other elements at
selected stations in the Southern Hemisphere. A new survey of this
work seemed in order, particularly as forecasting possibilities seemed
to include areas in which military operations were taking place, and
ten to twenty years of data were available for an independent test
of Walker's work.
Accordingly,
a small staff began work in January 1943, on the following problems,
selected because of their obvious importance in military planning
of future campaigns:
- Forecasting
the summer (monsoon) rainfall of Burma.
- Forecasting
the winter temperature and precipitation of Northwest Europe.
- Forecasting
the summer rainfall of Japan.
- Forecasting
the August to November rainfall of selected points in the East Indies
and Formosa
- Forecasting
the July and August rainfall of southern Italy and the Mediterranean
coast of France.
The program
called for a review of literature to find all work done previously
on the particular problem, and a search for variable features of the
general circulation of the atmosphere which might be expected to control
the variations in weather for the selected locations and seasons.
In general, an additional ten years of data were available since the
earlier work, and these data were used for a test of the continued
validity of such relationships as had been suggested in the published
literature. If this test looked promising, a forecasting formula was
developed, and experimental forecasts prepared.
This
work was carried on in part in cooperation with Blue Hill Observatory.
Although it was possible to develop a forecasting formula for each
of these problems, their reliability was not great enough to justify
a high degree of confidence in the forecast. Since it was not possible
to ascribe a clear-cut physical process to the relationships which
were used, little confidence could be placed in the formula as originally
derived and the period of record was in no case long enough to permit
a conclusive statistical evaluation of the formula on independent
data. It is believed that this method of approach to the long-range
forecast problem would be of considerably more value if the development
and testing of relationships could proceed from physical reasoning
to statistical testing in small units such that each step in the physical
process envisioned could be placed on the firm foundation of fact,
rather than left as a hypothesis.
4.
Forecasting From Solar Phenomena.-Two investigators,
Dr. C. G. Abbot, former Secretary of the Smithsonian Institution,
and Mr. H.H. Clayton, have from time to time reported on studies of
the relation of sunspot, solar radiation, and solar faculae variations
to subsequent weather variations. So little is known about the effects
which variations in solar radiation might have on the earth's atmosphere
that these studies have been almost wholly empirical.
Nevertheless,
it is generally recognized that variations in the solar output of
energy would in some manner be reflected in the atmospheric processes
on the earth, and it is entirely conceivable that variations in the
electrical or magnetic fields of the earth due to solar variations
other than in radiation might have important influences on weather.
Although
research funds available to the Weather Bureau did not permit thorough
investigation of these possibilities, some studies were made in an
attempt to check hypotheses suggested by Dr. Abbot and Mr. Clayton.
A breakdown was made according to years, seasons, latitudes and ocean
or land locations, but a detailed analysis of variance disclosed no
statistically significant relationship between the sunspot activity
as measured and the variations in the surface pressure. Even those
effects which approached statistical significance were too small to
be considered of practical importance in the immediate development
of long-range foreecasts.
5.
Forecasting From Ocean Currents and Polar Ice.-An
extensive investigation was made of the possibility of estimating
the variations in transport of the Gulf Stream as a basis for explaining
part of the variation in temperature at stations in Great Britain,
France, Germany and other countries of Northeast Europe. Earlier work
had been done on this problem but meteorologists were not in agreement,
nor were oceanographers, as to the extent to which variations in Gulf
Stream transport would influence surface water temperatures in the
northeast Atlantic, nor the extent to which such variations of water
temperature would influence air temperature over adjacent land areas.
The plan
in this study was to estimate changes in Gulf Stream transport by
using wind observations along the course of the Stream, on the assumption
that the greater the component of wind speed in the direction of flow,
the greater would be the volume transport. It is known that this assumption
is not entirely correct, but direct observations of Gulf Stream transport
were not available. The statistical analysis gave a slight suggestion
of a real relationship between the wind and subsequent temperatures
at lags of five and six months, but even those effects were too small
to be of practical importance for forecasting.
In cooperation
with Blue Hill Observatory, a study was made of the relation of the
extent of ice in the polar seas to subsequent temperature and precipitation
in Europe. Although this study was based on what seemed to be reasonable
physical theory, many cases were found in the record which did not
support the hypothesis. Nevertheless, an experimental forecast was
made for a number of European stations for October to December 1944.
In general the verification was poor. Work on this project was discontinued
in July 1945, because of the poor prospect of obtaining useful results
in the immediate future.
6.
Persistence as a Factor in Forecasting.-An
attack on the basic problem of persistence in the general circulation
was made by computing correlations between successive five-day pressure
patterns for the Northern Hemisphere stratified into areas representing
the branches of the general circulation. The object was to discover
under what conditions pressure patterns tend to be stable (persistent)
and under what conditions they tend to change rapidly. The period
for which five-day mean pressures were available was not long enough
for any conclusions to be drawn, although it appeared from the preliminary
study that there were results of potential value in extended and long-range
forecasting. This study should be continued with the more adequate
data now available.
Another
project was carried through to examine the persistence in monthly
mean temperature and precipitation. Auto-correlations with lags up
to 12 months were computed for 100 years of London temperature, Rome
temperature and precipitation, and New Haven and St. Louis temperature
observations. Significant correlations were found and in the case
of London were used in preparing experimental monthly temperature
forecasts for summer months. Because these forecasts did not have
a high degree of reliability, and because personnel were not available
to make the necessary meteorological studies to combine this information
with other factors, the investigation was discontinued.
One further
project was completed which was significant in the development of
the extended and long-range forecast research program, but which was
not aimed specifically at solving a forecast problems. This was a
project to test the potential usefulness of prognostic pressure patterns
in extended forecasting. The project was designed to help indicate
the direction in which research efforts might be most effectively
used to quickly increase the accuracy of extended forecasts. The results,
by indicating the relative magnitude of errors in two phases of the
forecasting procedure, led to a useful change in emphasis in the extended
forecast research program. Whenever in the future it is required that
research facilities be used in the most effective manner possible,
such experiments should be very useful in pointing the direction for
research and development projects.
Flight
Advisory Weather Service.-The number of
military flights (scheduled, training, and transient) increased steadily
throughout the period of the war. Protection of these flights, as
well as protection of civilian scheduled and transient operations
required the development of a meteorological service which would concentrate
its attention upon their specific needs.
To meet
these requirements the Weather Bureau issued to this filed stations
in June 1943 instructions concerning the provision of special services
to pilots and to traffic controllers. At the same time an experimental
unit was established at la Guardia Field to provide highly specialized
service to the Airway Traffic Controllers, and through them to the
pilots in flights. The success of this installation resulted in establishment
of another unit at Washington National Airport in September 1943,
this one planned along somewhat different lines. At the request of
the Army Air Forces this program was expanded as rapidly as possible,
and 22 additional units were established during the late months of
1943 and the early months of 1944.
At about
the same time the Army Air Forces inaugurated their Flight Service
Centers, located adjacent to the Flight Advisory Weather Service units
in the Air Traffic Control centers. Army Flight Service personnel
acted as clearance officers for many classes of military flights and
also checked every military flight plan to be sure that weather was
satisfactory and that navigational facilities and field conditions
were suitable.
The volume
of traffic handled increased steadily to a peak in July and August
of 1945 when more than 26,000 clearances per month were issued and
flight plans involving more than 650,000 aircraft per month were handled.
The meteorological information for this task was provided by the Flight
Advisory Weather Service. The Army has expressed full approval of
the service rendered by the Flight Advisory Weather Service units
and has indicated that only a very small number of advisories were
found to be unsound.
Personnel
were provided to maintain a continuos watch over weather along the
airways and at terminals. Teletype installations in the Flight Advisory
Weather Service offices insured receipt of airway forecasts and weather
reports as early as possible, and close contact with communications
operators provided prompt receipt of in-flight weather reports transmitted
by radio. With this information available the Flight Advisory Weather
Service forecaster was able to keep Air Traffic Control and Army Flight
Service personnel advised promptly of weather developments and anticipated
changes which were significant in the dispatching, control, and protection
of both civil and military aircraft.
-
Top of Page -
