1212 
signal-to-noise ratio at the receiver becomes unfavor- 
able. 
To describe the error in wind-velocity measurement, 
use will be made of the standard vector-error oy. This 
quantity is defined by the equation 
o, = Vo2 + | V ?o% (mph), (2) 
where a; is the standard wind-speed error, |V| the wind 
speed, and co, the standard wind-direction error. The 
standard vector error then represents the radius of 
the error circle associated with each wind-velocity 
measurement. Errors in speed and direction are thus 
conveniently combined. The measurement of o,, how- 
METEOROLOGICAL INSTRUMENTS 
sonde. It was assumed that the rate of rise of the 
balloon was 1000 ft min“, that the probable incre- 
mental error in pressure measurement was --1.0 mb 
(a very conservative estimate), and that the probable 
error in angular measurement was -£0.05 degrees. 
These data are sufficient for computing the standard 
vector error by equation (2). The numbers at the side 
of the wind-speed curve represent standard vector 
errors in miles per hour. Figure 7(A) represents an 
unfavorable situation for wind measurement, Fig. 7(B) 
a favorable one. In (A) very strong westerly winds 
were prevalent, reaching speeds of 100 mph at 40,000 
ft. The balloon burst at approximately 120,000 ft 
WIND DIRECTION 
O° 90° 180° 270° 360° 
a,c08 WIND 
60,000 
40,000 
20,000 
{0} 
0 10 20 30 40 50 60 70 80 90 IOO MPH (0) 
SLANT 
RANGE 
DIRECTION 
| 
O 10 20 30 40/50 60 (DIRECTION 
: al 
MILES , 
O° 90° 180° 270° 360° 
ok 
pe 
a 
10 20 30 40 
WIN e} 
We eS 
10 20 30 40 50 60 MPH 
(A) WIND SPEED (B) 
Fic. 7.—To compute the standard vector error expected from the radiosonde-radiowind set, two wind structures observed at 
Belmar, N.J., in December and July, are shown. In A, a strong jet is present at 40,000 ft, with the winds predominantly westerlies 
up to 120,000 ft. In B, the winds from 60,000 to 120,000 ft are strong easterlies. Computed standard vector errors are placed along- 
side selected height intervals of the wind-speed curves. 
ever, requires knowledge of azimuth elevation angle 
and height errors, and is influenced by the time in- 
terval used for averaging the wind speed |V|. With 
the present equipment, 2-min time intervals are used 
up to 40,000 ft, 5-min intervals up to 70,000 ft, and 
10-min intervals up to 100,000 ft. 
It is clear that to ascribe a single standard vector- 
error to the wind-measuring system is not meaningful 
unless considerable auxiliary knowledge is also avail- 
able. Sufficient information for vector-error computa- 
tion is given in the two graphs of Fig. 7. These graphs 
represent two wind cross-sections observed at Belmar, 
New Jersey, Fig. 7(A) during December, and Fig. 
7(B) during July. For ease in computation, only the 
gross features of this wind structure are shown. In 
each example the wind speed and direction are shown, 
together with the computed slant range of the radio- 
nearly 65 miles from the direction finder. The sudden 
decrease in error from 40,000 to 60,000 ft is due prin- 
cipally to the use of 5-min rather than 2-min averaging 
intervals. Above 80,000 ft, the errors in height deter- 
mination become critical. By a judicious use of the 
rate of rise of the balloon, this error can be reduced by 
a factor of two or three. 
In Fig. 7(B), the winds were from the northwest, 
increasing in magnitude to 40,000 ft, then decreasing 
to nearly zero at 60,000 ft. Above this level the winds 
steadily increased, blowing from the east and returning 
the radiosonde to the vicinity of its launching point. 
The balloon burst at approximately 120,000 ft. For 
this case the elevation angles for the high altitudes 
were in the vicinity of 80 degrees, and errors due to 
height imaccuracies did not predominate; hence the 
rather low standard vector-error above 80,000 ft. 
