752 
B. A positive deviation means that the cyclone moved 
to the right of the upper-level contour or isotherm. The 
mean deviation was reduced to 19° by averaging the 
direction of contours in the layer from 700 mb to 500 
mb with the direction at the 200-mb surface. Austin 
also found that a cyclone which was being steered 
tended to continue being steered, that the deviation 
appeared to be independent of intensity, that approxi- 
mately 65 per cent of the cyclones moved to the right 
of the upper-level isotherms and contours, that the 
difference between filling and deepening cyclones was 
negligible, and that the average deviation was greater 
for the Rocky Mountain plateau than for the eastern 
United States. Forecasters will agree with Austin’s con- 
clusions that the forecasting principle of steering com- 
pares favorably with other prognostic procedures for 
determining the direction of movement of a cyclone 
and, indeed, is the best, although definitely not a pre- 
cise tool, for this purpose. In this most useful study it 
did not appear that the speed of a cyclone could be de- 
termined from the geostrophic wind speed at any partic- 
ular level, and it is believed that the poor correlation 
obtained came as a surprise to most forecasters. 
In a somewhat similar study of cyclones and anti- 
cyclones, Longley [82] found the deviations given in 
Table III. 
Taste III. Mean Vatues ror Ciasses GroupeD ACCORDING 
To THE 700-mB FLow 
Mean | Mean 
Number| Mean | 24-hr | devi- 
Class of devi- | move- | ation 
cases | ation | ment /distance 
(mi) | (mi) 
Cyclones 
Closed centers aloft...........| 87 44° | 468 | 332 
Troughs and ridges aloft........ 36 20° | 537] 310 
Straight contours aloft......... AT 17° | 628 |] 295 
Anticyclones 
Closed centers aloft...........] 46 65° | 311) 295 
Troughs and ridges aloft........ 30 37° | 515 | 400 
Straight contours aloft.........| 33 27° | 587 | 328 
The frequency distribution of the deviation angle 
derived in Longley’s study, which, for cyclones, is in 
good agreement with Austin’s, appears in Table IV. 
TaBie IV. Frequency DisTRiBUTION OF THE DEVIATION 
ANGLE 
| 41802 to | +45°to | +15°to | —20°to | —s0° to 
+50° +20° —15° —45° —180° 
Anticyelones..... 14 13 38 20 24 
Cyclones......... 22 35 82 22 9 
Longley found that, of 47 anticyclones with a deviation 
of 45° or more, 33 were located over the Atlantic along 
a line approximately from 45°N and 30°W northeast- 
ward to 55°N and 20°W, and that a greater concen- 
tration of cyclones with large deviations existed in a 
rectangle bounded by the 40°N and 50°N parallels 
and the 50°W and 70°W meridians, or as the pressure 
systems approached the Atlantic high and the Iceland- 
Greenland low. Longley differed with Austin in regard 
WEATHER FORECASTING 
to the tendency of a cyclone to continue to be steered 
if following the upper-air flow but agreed that the wind 
velocity at the 700-mb surface was not a good indica- 
tion of rate of movement of the surface cyclone. 
Palmer [41] obtained as good or better results on di- 
rection of movement by a simple graphical device em- 
ploying the following parameters: 
1. Normal 24-hr direction of movement based on 
data published by Bowie and Weightman [7]; 
2. The direction in which the cyclone moved during 
the past 6 hr; 
3. The orientation of the major trough in the sea- 
level pressure pattern; and 
4. The direction from the 3-hr anallobaric center be- 
hind the cyclone to the 3-hr katallobaric center ahead 
of the cyclone. 
Tests indicated that this tool would be correct within 
15° about 75 per cent of the time. Palmer considered 
only winter storms; summer lows might not yield as 
good results. 
The procedures employed by both Austin and Long- 
ley perhaps do not provide a wholly fair test of the 
validity of the steering principle, but tests should be 
made of all forecasting techniques and rules which are 
held im high regard by all forecasters. In general, fore- 
casters simply do not know the accuracy of many of the 
rules and techniques in use or the range of the possible 
deviation of nearly all of them. 
In applying the results of Austin’s study, the fore- 
caster is faced with the determination of the probable 
direction of deviation from the steermg mdicated by the 
700- and 500-mb charts. 
The deeper the low, the less applicable the steermg 
principle. Surface lows which have closed centers at 
700 mb and higher tend to move with the upper center. 
The best indication of the direction of movement of 
closed lows at the 700-mb surface and higher is the 12- 
hr pressure change around the center; another method 
is the determination of the resultant of the wind cir- 
culation around the upper-level low. The calculation of 
the distance such a low will move im a given time is 
even more difficult and should be determined by the 
broad-scale weather processes going on. 
Little has been written on the practical applications 
of the 200-mb chart to forecasting. Wulf and Obloy 
[59] have emphasized the importance of the compensat- 
ing effects between the stratosphere and the lower 
troposphere and have suggested that stratospheric con- 
ditions may exert considerable influence upon fronto- 
genesis and that advection and other processes in the 
lower stratosphere should receive consideration in prog- 
nosticating surface and tropospheric pressure patterns. 
Austin [1] found that the 200-mb chart is almost as 
effective a steering level as any other. Some forecasters 
have found the 200-mb chart helpful in determining 
displacement of shallow migratory troughs at the 
700-mb level. 
Under the tropopause, at approximately the 300-mb 
level, an extremely fast and narrow current has been 
noted which is called the ‘jet stream” by the University 
of Chicago group [53]. Large-scale mixing, interrupted 
