WMOD00484

• Title: Microphysical Effects of Wintertime Cloud Seeding with Silver Iodide Over the Rocky Mountains - Part I
• Date: 10/10/1988
• Weather Modification - Doc Type: Report

1148

JOURNAL OF APPLIED METEOROLOGY

VOLUME 27

time required to cycle the unit to calculate the SL W
content, for 17 cases having continuous liquid water
over the distance required for one cycle of the probe.
Droplet collection efficiency was considered to be unity
at typical aircraft flight speeds. Cycle times ranged from
11-70 s and resulting mean SL W contents were from
0.22-0.03 g m -3. These were compared with the mean
J-W contents for the same time intervals. The corre-
lation coefficient for the 17 data pairs was 0.975. The
J-W tended to be about 0.01 g m -3 higher for values
near 0.05 g m -3, while the Rosemount tended to be
about 0.03 g m -3 higher for values near 0.15 g m -3 .
This comparison indicates that the J-W SLW content
values were reasonably accurate.

d. AgI detection

An NCAR acoustical ice nucleus counter was used
to determine the approximate AgI plume position in
each experiment. The acoustical counter will be dis-
cussed in some detail because it is not as commonly
used as other aircraft instrumentation. Although a
number of individuals or groups have had inconsistent
results with acoustical counters, properly maintained
and operated systems have provided consistent results
in detection of AgI from day to day and location to
location. For examples, see Super (1974) and Heim-
bach and Stone (1984).
Different but very similar counters were used on the
two aircraft. They were improved versions of the coun-
ter discussed in detail by Langer (1973). Air from out-
side the aircraft cabin was sampled at about 10 L min-1
and had water vapor and cloud condensation nuclei
added prior to injection into a refrigerated 17 L cloud
chamber maintained near - 20oC. This resulted in
cloud formation near the chamber top with droplet
concentrations well in excess of natural cloud. The high
droplet concentration was intended to limit transient
supersaturations to a few tenths of one percent and to
partially offset the short chamber residence time (1-2
min for most particles), which is especially important
for ice crystal formation that is due to contact nucle-
ation. The work of Demott et al. (1983) suggests con-
tact nucleation as the predominate mechanism with
the type of seeding material used. In spite of the marked
differences between natural clouds and those main-
tained in the acoustical counter, Langer and Garvey
(1980) reported good agreement between measure-
ments of ice nucleus concentrations with various AgI
smokes using an acoustical counter and measurements
obtained in the Colorado State University (CSU) iso-
thermal chamber. The latter simulates natural clouds
much better than the former.
Cloudy air was continuously drawn out the bottom
of the counter chamber through an acoustic sensor,
where ice crystals that had grown larger than about 20
~m produced audible clicks. These sounds were con-
verted to voltages by a small microphone and the total

"counts" then processed and recorded once per second
by appropriate electronics. The formation of large
droplets in the cloud chamber injection tube, which
could also trigger the sensor,. was avoided by slightly
heating the tube to prevent condensation.
Laboratory work has revealed that only about 10%
of the ice crystals formed in the cloud chamber reach
the acoustic sensor due to losses on the glycol-covered
walls and exit cone (Langer 1973). However, unless
otherwise noted, the data reported are simply the counts
as detected by the acoustic sensor.
The acoustical counters, operated at - 200C at both
experimental sites, indicated that the background con-
centration of natural ice nuclei was almost zero.
Whether numerically correct or not, this characteristic
made detection of AgI straightforward, since experience
showed that encounters of more than a few counts per
minute were always downwind of an AgI source. As
used in these experiments, the acoustical counters were
essentially AgI detectors rather than natural ice nucleus
detectors, and their response will be presumed to be
due to AgI only.
The entry edge position of an AgI plume can be
approximated by estimating the delay time to first
counts. The width of a ground-released plume can be
estimated by making pairs of passes in opposite cross-
wind directions as done by Super (1974), which as-
sumes that plume meander is not significant in the
several minutes between passes. This is not always
valid, particularly in very stable conditions.
Each counter's delay time was empirically deter-
mined from ground tests in which varying amounts of
AgI smoke were injected into the counter by hypoder-
mic syringe over a few to several seconds. The delay
time (.:It) was defined as the first second of the first
seven-second period having three or more total counts.
Because nucleation, growth and detection in the
acoustical counter is a stochastic process, first-response
time was well related to the total counts ( C) per injec-
tion by a power curve of the form

.:It = AC-P,

(1)

where A and P were empirically determined constants.
As an example, a correlation coefficient of 0.88 resulted
from this power curve fit to the data from the counter
used over the Grand Mesa, where A = 50.6 and P
= 0.10. Individual data points were within :t3 s of the
curve for total counts greater than 100, but some de-
partures were as great as 6 s for lower amounts of AgI.
Equation (1) was used to estimate the entry edge po-
sition on each flight through an AgI plume where C
was the total counts recorded during the pass. However,
(1) does not strictly apply because the time to traverse
a plume was usually much longer than the AgI injection
period in the ground tests. As a consequence, derived
plume widths are narrower than the actual widths. The
problem is not usually serious because, for example,