WMOD00510

• Title: The Practicability of Dry Ice for On-Top Seeding of Convective Clouds
• Date: 1/1/1978
• Weather Modification - Doc Type: Report

.

,

\
~
r

.

1

I.

I

JANUARY 1978

HOLROYD, SUPER AND SILVERMAN

53

ity of the foil. No' 50 ~m impressions were resolved.
tItany event the experiment showed that the CSIRO
foil impactor, with careful illumination of the foil, can
detect column crystals down to about 100 ~m lengths;
however, the counting effici.ency is unknown.

d. Case studies of clouds

Several summertime clouds were selected for experi-
ments to show the productivity of ice crystals by dry
ice. In the American experiments near Miles City,
Mont., each cloud was penetrated before seeding to
measure natural ice crystal concentrations. The cloud
was then seeded by a separate aircraft dropping dry
ice pellets from just above or below the cloud top.
In the Australian experiments only one aircraft was
used near Emerald, Queensland, to seed, the cloud on
the first pass and then look for crystals after the danger
of riming the foil impactor had lessened.
In nearly all cases a cloud was selected that was
reasonably isolated from others so that it could be
identified for repenetration. When selected, the clouds
had no visible glaciation and had firm boundaries.
Subsequent photo analysis showed that many tops
were descending at the time of seeding. The American
clouds are' assumed to be continental in nature. Pre-
liminary cloud droplet and CCN measurements tend
to confirm this. The Australian clouds at 230S latitude
and about 125 km from the coast are assumed to be

o

c.-I'-
'Z
1010
u_
It:
101
a..

10

.7

FIG. 2. The spectrum of column lengths measured on 25 JLm
thick aluminum foil. Part (a) shows a peak frequency of 250 JLm
for the sample size of 829 crystals. The cumulative distribution is
plotted on log-probability axes in part (b); the nearly lognormal
distribution shows no evidence of truncation at sizes, 100 JLm and
larger.

I~,' -1" "

~i:~_;;r:~-'_i :',',:.: ',,;_: _ ,,'f ';., ~~_,,:, _ "\/ .:'_ _ : ''1.-,_,_,[ i ,..'>,- -'_~~" '_ :',...;:,,::;..:-,,;ii,: (.t';~"H<~ '/_".'2~
~., .iIi-.""idi""",c.,."~":,;"";"'_"'''''''',j.~,,,,,~..,,.,~~~...~.,.;

mostly modified maritime tropical in nature, though
no supporting measurements were made. On many
days there appeared to be rapid ice multiplication at
relatively warm temperatures in the Australian clouds.
The clouds were presumably satisfying the Hallet and
Mossop (1974) ice multiplication conditions of large
droplets, graupel, and a top-dwelling at -S to -lOoC
for a while. Clouds were not seeded when rapid glaciation
was visually in evidence in neighboring natural clouds.
This still does not rule out the possibility that the
crystals sampled in the Australian clouds resulted
from causes other than seeding.

3. Data analysis

The glaciated cloud volume, typical crystal concen-
tration and mass of dry ice used need to be measured or
, estimated to determine the effectiveness of dry ice in
creating ice crystals.
The cloud volume dimensions were determined from
the flight track and photography. The path length of a
penetration is known from integrating the true air
speed. Using the distance to a cloud, known from the
flight track, photographs of the cloud were measured
to give horizontal and vertical dimensions. A local
sounding by aircraft or rawinsonde was used to obtain
the altitude of the OoC level. In this way many measure-
ments were available ,to define, with some probable
error, the supercooled cloud volume. This volume was
considered to be a rectilinear solid with length and
width determined by track lengths and by photos, and
thickness determined by height above the OoC level
and by photos.
The crystal concentrations were determined by the
instruments described above. To make some allowances
for nOllluniformity of crystal concentration a linear
average of these concentrations across the entire cloud
was calculated for each pass. When many passes were
made, those selected as representative were at times
late enough to allow the crystals to spread appreciably
throughout the volume, yet early enough so that the
majority of crystals had not fallen to levels below the
sampling level. It was assumed that the crystals
filled the vertical cloud dimension from OoC to cloud
top and that the natural background concentration of
ice crystals did not increase during these experiments;
these may not always be good assumptions. However,
in the case described in detail below it was verified
that the crystals did distribute themselves horizontally
and vertically throughout the described volume.
The effectiveness of dry ice is given by

LWHC
E=-
M

(5)

for cloud length L, width W, thickness H, crystal
concentration C and CO2 mass M. These values are
given for each experiment in Table 4.