1278 
sense. Fair-to-good relative correlations have been found. 
Figure 16 shows pulse-integrator records of two radars 
(with operating wave lengths of 3 and 10 cm) compared 
with a rainfall intensity record obtained from a modified 
Fergusson rain gage located 12 miles away. The two 
sets were directed and ranged over the gage. Fluctua- 
tions of the signal intensity of both radars show good 
agreement with variations of the rainfall intensity. The 
differences between the two radar pulse-integrator 
curves are due to differences between the radars (fre- 
quency, beam width, attenuation, and pulse length), 
and perhaps to slight errors in orientation and ranging. 
Radar Signal Spectrograph. It has been shown both 
by mathematical analysis and by observation [80; 65, 
p. 35] that relative motions of scattering particles in 
range establish audio and sub-audio frequency fluctua- 
tions in the power return of radar storm echo signals. 
This is caused by the changing phase relationship, or 
interference patterns, of the return radiation from the 
separate particles as they move with respect to one 
another. This relative motion may take place in at least 
four different ways: 
1. Random motion of the scattering particles. 
2. Differences in fall velocity of the particles. 
3. Small-scale turbulence. 
4. Horizontal air streams moving in different direc- 
tions or with different velocities, or both. 
One type of radar signal spectrograph—designated 
by the abbreviation “rasaph’’—is an electronic instru- 
ment designed to scan a range of frequencies from 3 to 
300 cps and to record on a meter the average power of 
echo signals at any interval within this range. If the 
relative motion of the particles is random, the peak 
signal intensity will lie at zero cycles. As a great deal 
of random relative motion always exists, the spectro- 
graph trace invariably shows a pronounced peak at the 
lowest frequencies observable. If large numbers of hy- 
drometeors are moving at some definite velocity with 
respect to another large group (for example above and 
below a shear zone), a definite frequency is established 
in accordance with the following relation: 
y = 2u/X, 
where v = fluctuation frequency, uw = difference in ve- 
locity (along a given azimuth from the radar) between 
the two groups, and \ = operating wave length of the 
radar. As is apparent from this relationship, high rela- 
tive velocities between particles can cause high fluctua- 
tion frequencies. However, with pulsed radar there is a 
limit to the maximum frequency that can be observed; 
frequencies higher than half the pulse repetition fre- 
quency of a particular radar cannot be observed by 
that radar. (Pulse repetition frequencies usually le be- 
tween 200 and 1000 per second.) 
Only relative, not absolute, velocities may be deter- 
mined from the recordings of this instrument. Not 
enough observations have been made at this writing 
to enable accurate evaluation of its usefulness, but the 
following are a few potential uses which it may have: 
1. Determination of turbulent conditions in storms. 
RADIOMETEOROLOGY 
2. Determination of the distribution of fall velocity 
for various kinds of hydrometeors. From this, the drop- 
size distribution may be obtained. 
3. Determination of the behavior of freely falling 
particles as they descend through regions of wind shear. 
4. Information concerning the horizontal velocity 
distribution of hydrometeors may be obtained if wind 
velocities are known from rawin observations. 
Storm Echo Signal Contouring. Harly investigators 
in the field of radar storm detection learned that while 
no simple relationship between rainfall intensity and 
echo-signal power existed, regions of more intense pre- 
cipitation in a given storm could be located with fair 
accuracy by reduction of receiver ‘‘gain’’ to a pomt 
where only the strongest echo signals showed on the 
PPI scope [36]. By plotting the outlime of the storm 
between successive equal gain reductions, the storm 
echo signal was reduced to a number of contours; each 
contour represented a level of equal echo-signal power. 
In general, it was found that the strongest echo signals, 
which were loosely associated with more intense rain- 
fall, were near the center of storm-cell echo signals. It 
was then suggested [5] that the radar be made to do 
the work of contouring, and to accomplish this in a 
coarse fashion it was necessary only to block the echo 
signals of greater than certain power so that they would 
not appear on the PPI or RHI scopes. Then the area 
of strongest echo-signal power remained dark in the 
center of the storm cell presentation. To verify the 
presence of precipitation in this area, the blocking cir- 
cuit could be shut off and the scope then regained its 
normal appearance. Only two levels of signal power 
could be discriminated by this technique, but this was 
sufficient for initial tests. 
The American Airlines System tested the usefulness 
of this technique for aircraft storm avoidance, with in- 
teresting results [11]. An experimental test aircraft was 
equipped with an X-band radar modified to produce 
storm contouring when desired. This plane was delib- 
erately flown into thunderstorms to test the reliability 
of this equipment for detection of areas of heavy tur- 
bulence and precipitation. It was found, at least for the 
storms penetrated, that areas of heavy turbulence did 
not necessarily coincide with regions of intense echo 
signal, as indicated by the contouring. Turbulent regions 
tended rather to lie where the radar indicated that 
steep gradients of rainfall intensity existed, as shown by 
narrow, closely packed contours. The contouring proved 
quite reliable for determining areas of heavy precipita- 
tion. 
One difficulty with this system is contour distortion 
by precipitation attenuation, although this is not con- 
sidered very serious. When X-band (3-cm) radars are 
used, this is a factor to be considered, because of the 
heavy rainfall that occurs in most thunderstorms. Addi- 
tional distortion occurs as the aircraft approaches the 
storm, because the echo-signal power from the nearest 
particles increases more rapidly than from those on the 
far side of the storm, as a result of the range attenuation 
effect. Studies of these distortions are in progress in 
