720 
Coordinated Use of Sea-Level and Upper-Air Analysis 
Having once arrived at a satisfactory analysis of both 
sea-level and upper-level charts, we come to what is 
perhaps the greatest problem facing the forecaster 
today: the coordination of the sea-level data with the 
upper-air data. Most meteorologists are now familiar 
with both the sea-level and the upper-level charts, and 
routinely take both into account before rendering judg- 
ments upon the development and motion of storm 
centers and their associated manifestations of weather. 
Despite this fact there still appears to be an appreciable 
difference between the scope of the synoptic meteor- 
ologist’s general knowledge of weather processes and 
the part of this information which he actually applies in 
analysis or forecasting. It appears that many forecasts 
miss their mark not because of our over-all lack of 
understanding of the factors which control the changes 
in the weather, but because the coordination of all the 
various pertinent details at all levels of the atmosphere 
in a relatively short time interval is a task too difficult 
to accomplish satisfactorily with our present system of 
map display and data representation. 
If we try to picture what is needed in order to co- 
ordinate with facility the meteorologically significant 
features of the higher strata of the atmosphere with 
those at the surface, we shall arrive at something like 
the following: 
First and most important, all the various charts de- 
picting conditions at the surface and at upper levels 
should be the same size. This proviso, though simple, is 
vital if the forecaster is to be able to compare one level 
with another quickly and accurately. 
Second, the charts for all levels should be drawn on 
semitransparent paper in order that one may be super- 
imposed upon that for an adjacent level to facilitate 
examination of the changes of meteorological features 
with height and to insure internal consistency im anal- 
ysis. Furthermore, it is obvious that we should use the 
same units for data at all levels if the maximum ease in 
vertical synthesis is to be obtamed. Consider, for in- 
stance, the disordered array of units in current use: 
centigrade and Fahrenheit temperatures; Beaufort 
scale, knots, miles per hour, and meters per second 
denoting wind velocities; altitudes measured in feet, 
meters, or dynamic meters; and pressure analysis 
hampered by the use of isobars on sea-level maps and 
contour lines on upper-level charts. The system of units 
proposed by Bellamy [2] is of a unitary type. As has 
been mentioned, he has also proposed a system for the 
graphical representation of data, which obviates the 
need for converting numerical values from one system 
of units to another. By means of simple transparent over- 
lays, the numerical values of the data in any desired 
units may be read directly from the graphs. 
Although complete standardization of all units 1s not 
feasible at this time, we may even now progress towards 
this goal by adopting a single system of units for pres- 
sure. In particular, the unfortunate duality of units with 
isobars drawn on the surface map and contour lines on 
the upper-level charts can be abolished without com- 
plexity by replacing the sea-level isobars with the 
OBSERVATIONS AND ANALYSIS 
1000-mb contours; the observational stations could re- 
port the height of the 1000-mb surface as well as the sea- 
level pressure. 
At present, in our attempts at vertical synthesis of 
atmospheric elements, we labor under a difficulty other 
than that occasioned by a multiplicity of units; we 
refer to the unfortunate time interval which elapses 
between surface and upper-air observations. As a result 
of this lack of synchronization, many of the fundamental 
charts upon which analyses and forecasts are based are, 
in effect, mvalid. Rectification of the observational 
program must inevitably precede sound coordination of 
upper-level charts with surface data. 
A further argument in favor of using uniform map 
scales and units plus transparent or semitransparent 
map bases is that they make possible the construction 
of a chart which we believe to be most effective in 
bringing out the relationship between surface and upper- 
air circulations. This is simply the sea-level chart with 
upper-level contours traced on it, preferably using for 
the upper-level contours some color differmg from that 
of the sea-level isobars. This method seems to be the 
most satisfactory for relating the patterns at upper 
levels with the weather and pressure changes occurring 
at sea level [13]. Such a chart can be used to especial 
advantage by those who must issue forecasts or com- 
plete an analysis of the synoptic situation in a very 
limited time. When time is pressing, analytical compre- 
hensiveness 1s all too frequently sacrificed as long as the 
surface and upper-level charts are kept separate, for 
even when both are carefully studied independently, 
the possible inferences to be drawn from the changes of 
flow with height can be assimilated only by a careful 
comparison which consumes more time than is now 
available to most operational meteorologists. 
If this general concept for the coordination of surface 
and upper-air data is accepted up to this point, the 
problem then arises as to the selection of the upper-level 
chart most productive from the standpoimt of the num- 
ber of utile inferences deducible from the combined 
analysis. This set of upper-level contours depends to 
some extent on the primary purpose to which the charts 
will be put. If we consider that forecasting the appear- 
ance of the sea-level and upper-level pressure charts 
24—48 hours in the future is the primary problem, then 
we should choose for analysis the upper-level chart 
considered to be most significant for the investigation of 
steering, cyclogenesis, and anticyclogenesis. At present 
many meteorologists prefer the 500-mb chart over all 
others for this purpose. It would therefore be proper to 
superimpose on the surface map the 500-mb contours at 
analysis centers whenever pressure-pattern prognoses 
constitute the chief forecast problem. 
On the other hand, if the primary problem is to fore- 
cast the cloudiness and precipitation for the coming 
twenty-four to forty-eight hours, then the upper-level 
chart chosen should be the one best suited for the detec- 
tion of advection and the motion of the moisture layers 
which are most intimately associated with the observed 
cloudiness and precipitation. 
The problem of such short-range precipitation fore- 
