"Experimental Meteorology," V.J. Schaefer, 1950

Separata copy
JOURNAL OF APPLIED MATHEMATICS AND PHYSICS (ZAMP)

Vol. I, 1950 BIRKHAUSER PUBLISHERS, BASEL/SWITZERLAND Fasc. 3and4

Experimental Meteorology’)
Survey article

By Vincent J. Scuarrer, Schenectady, New York?)

As basic research along the frontiers of science advances, there are few of
the natural sciences which do not receive careful scrutiny. As each study ex-
pands, it encounters the borderlines of related phenomena which often become
so completely inter-related that it is difficult to determine where one ends
and the other begins.

This is particularly true when the study involves a subject such as the
earth’s atmosphere and the physical processes which occur within a few hundred
miles of its surface.

The general subject I should like to discuss at this time involves an area
of even shallower depth with few, if any, of the problems to be considered
involving a region exceeding ten miles above the earth. Despite this relatively
narrow zone in the physical dimensions of the earth, the problems encountered
are of considerable magnitude and of a nature which, if we are to approach
their solution with any hope of success, must employ as many scientific tech-
niques as possible.

As in any pioneering venture, success is much more likely if the investi-
gations are carried out with enthusiasm, imagination, and an active curiosity
and without too much regard for older theories or prejudices.

The role of clouds in the hydrologic cycle

In considering clouds and their formation due to the interactions between
air, water, and sunlight, we have the essential constituents of an important
mechanism for the release of energy on the earth. The water molecule is the
basic unit in this energy transfer system, since the water involved is con-
tinually passing through the closed cycle of evaporation, condensation, and
precipitation—the sequence starting again by evaporation.

Because of its physical nature, the lower atmosphere serves as a vast reser-
voir for water. A relatively high concentration of water vapor may be stored
in the form of gaseous molecules before condensation forms. In fact, even

1) This paper has been approved by the National Military Establishment and also by
R. W. Sarow (August 15, 1949). It has been presented before United Nations Conference on the
Conservation and Utilization of Resources (August 19, 1949).

2) General Electric Research Laboratory.

154 Vincent J. SCHAEFER ZAMP.

when this occurs, the small size of the condensed droplets often permits the
atmosphere to support this additional moisture for a considerable time—often
long enough, in fact, for the cloud thus formed to evaporate into warmer or
drier air. It is in this manner that large air masses become modified. Warm
and moist maritime air encountering colder, drier continental air mixes with
it at the contact interface and thus changes the nature of both air masses.

In addition to forming clouds, water molecules in the very low levels of
the atmosphere often condense as frost or dew. This is really a precipitate
but, except in regions where it is common, is not of great economic importance
since much of it soon evaporates again. Where cloudless skies occur and dew
deposits are common, they may constitute an important source of moisture
on the earth. Improved methods for enhancing this precipitation by artificial
means is an important problem that should receive more attention.

The presence ot clouds in the atmosphere tends to decrease the flow of
energy from the sun to the earth, since most clouds absorb much of the visible
and near infrared radiation of the sun. Even a thin layer of clouds reduces
insolation (incoming solar radiation). At night, however, the presence of clouds
provides an insulation blanket and slows down the loss of heat from the earth
to outer space.

When a cloud forms in the atmosphere, the amount of heat released to the
air amounts to about 580 calories per gram of condensed water. This energy
may be considered lost insofar as the water resources of the earth are con-
cerned if the cloud evaporates before it is precipitated. Not until precipitation
develops in the clouds as snow or rain and then falls to the earth may we
consider the hydrologic cycle completed and the maximum energy recovered.
Extremely complex relationships in this respect result from the interaction
between the seasonal changes of temperature and humidity, variations in the
wind, the effect of topography, land use, and similar factors, all of which
exert an influence on the delicate balance existing between the sun, the atmo-
sphere, water, and man’s welfare.

In discussing experimental meteorology, I would like to limit my consid-
eration to experimental studies in the lower ten miles of the atmosphere
and with air samples ranging in volume from a few cubic millimeters to several
hundred cubic miles.

Among the interesting problems related to the atmosphere which require
a solution is a logical and comprehensive understanding of precipitation pro-
cesses. That there is more than a single mechanism involved in the formation
of snow and rain as precipitation is now becoming well recognized. It was
only a few years ago, however, when heavy rain was only thought of {1]!) as
necessarily preceded by the initial formation of snow.

1) Numbers in brackets refer to References, p. 227, ZAMP I/4

Vol. I 1950 Experimental Meteorology 155

It seems to me that when we consider the economic aspects of precipita-
tion processes, it is important to evaluate not only the quantity of moisture
that reaches the earth but also the related effects, such as winds, lightning,
hail, intensity of the precipitation, and the related amounts of sunshine.

Tf any of these factors can be modified artificially, even to a limited degree,
it is of great importance that every feature of the modification should bear
careful scientific scrutiny.

While it is true that it is relatively easy to assign a definite value to an
inch of rain on a square mile of newly planted seed corn; ten thousand acre
feet of water added to an irrigation reservoir; two feet of snow on the upper
drainage basin of a stream used for power purposes; or to supply a large
metropolitan district with adequate drinking water, it is nevertheless of great
economic importance to evaluate weather phenomena. The loss in timber
resources and recreational areas occasioned by fires started by lightning storms;
the devastation resulting when high local winds flatten banana plantations;
the bruised fruit or smashed vegetables following an intense hail storm, and
the muddied streams and eroded farm lands left in the wake of a torrential
rain all serve as examples of the by-products of unstable cloud systems which
occur in the atmosphere.

Most cloud systems are examples of colloidal instability. Many of them
rigorously follow the reactions which have well known counterparts in colloid
chemistry. In most instances, where colloidal instability occurs, it is possible
to shift the system to a more stable form. by the proper and judicious use of
chemical or physical reagents or reactions.

Even though glycerine and boric acid, for example, can exist as super-
cooled liquids in a relatively stable condition, methods are known whereby
they can be changed to their less common, though more stable, crystalline
form. The water in cloud systems also has a tendency to develop supercooling
as a common characteristi

Cloud types to be considered

Before going into a detailed description of this interesting phenomenon,
it will simplify the discussion to establish a certain terminology at this time.
The clouds which we shall refer to consist of the three common forms—cirrus,
cumulus, and stratus. The many variations in which each of these may be clas-
sified will, in general, be neglected for, although they are very important from
the standpoint of the genetic development of a weather pattern, these variations
are not of prime concern in relation to the subject under consideration. A phy-
sical feature of great importance, however, is the freezing level in the atmo-
sphere and its relationship to the clouds under discussion. In general, the
temperature in clouds follows the wet adiabatic lapse rate with the temperature
156 Vincent J. SCHAEFER ZAMP.

lowering with increase of altitude at the rate of about 21°C (38°F) per
1000 feet at 0° C (32° F). With relation to the freezing level, we shall be deal-
ing with clouds existing under one of three temperature conditions. The term
waym cloud shall be used to designate a cloud of liquid water droplets warmer
than 0° C (32°F). A cold cloud shall then designate a cloud of liquid water
droplets colder than 0° C, while a cool cloud will be a cloud extending above
and below the freezing isotherm, thus combining the features of warm and
cold clouds. The term cool cloud may thus designate a cumulus cloud, cooling
according to the wet adiabatic lapse rate and thus becoming colder with
increase of altitude, or it may be used to describe a more stable stratus cloud
extending through an inversion and thus inverting the normal temperature
sequence. These various cloud conditions are illustrated in Fig. 1.

The amount of condensed water in clouds

The variation in the amount of condensed water vapor in natural clouds
covers a considerable range. In general, it is lowest in cirrus and highest in
cumulus, with stratiform types intermediate. The temperature of the source
air is the major factor governing the limit to the amount of moisture which may
appear in clouds from condensation. Table 1 (p. 176) shows the amount of gaseous
water vapor which may be contained in saturated air of different temperatures.

This difference in the amount of water which may be held as a gaseous
vapor at different temperatures is the direct cause of cloud formation in air
cooled by radiation, convection, or advection.

The size and size distribution of the cloud droplets contained within such
clouds is also a variable involving many complex relationships. Cloud studies
have now progressed to the point [2] where limiting values may be given in
sufficient detail to be quite satisfactory. These are assembled graphically in
Fig. 2. _

As might be expected, the lowest values in condensed water occur in cirrus
clouds as ice crystals, the intermediate values in stratiform clouds, while the
highest liquid water contents and particle sizes are found in towering cumulus
type clouds.

The degree of turbulence in clouds

Studies of turbulence and convection in clouds show that of the three
general types under discussion in this paper, only the cumulus type exhibit
turbulence and convection of high order. While it is true that vertical velo-
cities of a considerable degree may sometimes occur in stratus clouds sub-
jected to orographic or frontal lifting, i. e., displacement due to the encounter
of the cloud with a land barrier or an air mass having a different density, such
effects are of a temporary nature and generally fail to produce marked changes

Vol. 1, 1950 Experimental Meteorology 157

in the nature of the cloud. It is likely that vertical velocities in stratus clouds
rarely exceed 2 m/s. The vertical velocities in cirrus clouds are even lower,
since they generally form at very high altitudes and in relatively stable air.
Cumulus clouds, on the other hand, contain high velocities which may exceed
20 m/s in towering cumulus. Even in relatively small cumulus, 1 km in
thickness, vertical velocities of more than 3 m/s often occur.

Liquid water content and particle size are physical features of clouds which
are not affected by the location of the freezing level, except as they may be
a function of the absolute humidity. High values of particle size and liquid
water content may occur at temperatures far below the freezing temperature.
The factor which affects the liquid water content of a particular parcel of air
is the initial saturation temperature and its final temperature. If, however,
in reaching this final temperature, the cloud particles have grown so large
that the effect of gravity on them is great enough to overcome the vertical
air velocity of their surroundings, the liquid water content may become lower.
By a somewhat similar process, the liquid water content of a given air sample
may be temporarily enriched by an invasion of falling precipitation from
depleted clouds above. These are features of old clouds and are of much im-
portance in reaching a proper understanding of precipitation processes.

Types of nuclei in the atmosphere

In understanding the structure of natural clouds, it is of much importance
to consider the initial formation of the cloud particles. All of the features of
this phase of cloud physics are not understood, although rapid advances are
underway. Three types of nuclei are of importance—condensation, sublimation,
and freezing.

Condensation nuclei

If ordinary atmospheric air is saturated with water vapor and then suddenly
cooled, a cloud appears consisting of small water droplets. The number of
droplets which form is directly related to the effective condensation nuclei
present in the air before cooling occurred. If these observations are made using
an enclosed chamber, and its temperature and humidity are adjusted so that
the cloud droplets reach the bottom of the chamber before they evaporate,
it is an observable fact that successive expansion cycles fail to produce a new
cloud unless a higher expansion ratio is used at which time effective nuclei
serve as centers of condensation.

It a well known fact [3] that condensation nuclei are formed in great
quantities by many processes. Fine, hygroscopic salt particles, which become
airborne as waves and bubbles break at sea, seem to be an important source
of very active condensation nuclei. The smoke from forest fires, and most

158 Vincent J. SCHAEFER ZAMP

other burning processes, produces vast quantities of condensation nuclei which
permeate the atmosphere to a thickness in excess of a mile. In some industrial
regions, these particles become so numerous that they form a dense pall of
smoke and fog which restricts visibility to a fraction of a mile.

If air samples are used from industrial regions or other places where the
foreign particle concentration in the air is high, a dense cloud is observed
in the cloud chamber containing from 104 to 10° particles per cubic centimeter.
On mountain tops, at sea, or at higher levels of the atmosphere, the number
of effective nuclei may drop to values of oily a few hundred per cubic centi-
meter. It can easily be shown that the preponderance of such particles continue
to serve as water drop nuclei to a temperature of —38-5°C. Supercooled
clouds of this temperature can easily be formed, even when the concentration
of nuclei is as high as 105/cm%.

Sublimation nuclei

‘Among the foreign particles carried in the dust of the air are certain special
forms which in a proper environment serve as sublimation nuclei. These become
active in the formation of ice crystals at specific temperatures below 0° C
and in air supersaturated with respect to ice but not necessarily saturated
with respect to water. Fig. 3 shows typical results observable with samples
of natural soils and similar materials. Many of these were gathered in regions
where extensive dust storms are common occurrences.

It is perhaps of considerable significance that very few particles have been
found which serve as effective sublimation nuclei at temperatures warmer
than about —12° C. No observational evidence is known of snow storms start-
ing in clouds warmer than this temperature.

‘A study of considerable significance in relation to the occurrence of subli-
mation nuclei in the natural atmosphere has been made on Mt. Washington
during the past eighteen months. Every three hours, day and night, in con-
junction with the regular weather observations, the number of sublimation nuclei
in a typical air sample are determined as part of our Project Cirrus weather
research studies. The cold chamber method [4] is used in making these observa-
tions. The results are summarized in Table 2 (p.176). This indicates that the
level of sublimation nuclei in the atmosphere is generally very low throughout
the year. The highest level observed during approximately 4500 observations
is 10 per cubic centimeter. In contrast, the level of condensation nuclei at the
same time would show values ranging from 10 to 10,000 times greater.

While sublimation nuclei in the natural atmosphere may apparently have
a number of different molecular properties, there is one pure chemical com-
pound which to date has not been equalled in its effectiveness as a foreign
particle ice nucleus. This is silver iodide, a hexagonal crystal which is almost
identical in its crystalline structure to that of ice.

Vol. 1, 1950 Experimental Meteorology 159

Work in our laboratory [5] and in the field is actively underway evaluating
the effectiveness of silver iodide as a sublimation nucleus under various atmo-
spheric conditions. To date, it appears to have exceptional value for cloud
modification where dry ice cannot be used. One of its most important appli-
cations may be to inoculate clouds from ground generators placed in such a
manner that the submicroscopic particles are carried up into the clouds where
they will become active at temperatures colder than —5°C, the threshold of
activity of silver iodide smokes. Further details in relation to silver iodide
will be given in a subsequent section of this paper.

Freezing nuclei

Only brief mention will be made of freezing nuclei at this time, since not
much is known about them. It is apparent, however, that they seem to possess
somewhat different properties than sublimation nuclei. Where the latter per-
mit the formation of ice crystals by the direct deposition of water molecules
from the vapor to the solid phase, freezing nuclei appear to initiate the freez-
ing of supercooled water droplets [6]. It is apparently due to the presence of
freezing nuclei in bulk water that leads to the crystallization of water in the
temperature range of —6° C to —20°C under conditions when care is exer-
cised to prevent the seeding of the water by frost crystals deposited just above
the water meniscus on the container. It might be that the presence of certain
water insoluble sublimation nuclei are partially responsible for the develop-
ment of snow in cold and cool clouds at temperatures below —12° C when
they come in contact with supercooled cloud droplets and cause them to freeze.
It is a matter of observation that a considerable number of snow crystals
have as a nucleus what appears to be a cloud droplet such as shown in Fig. 4

Causes of natural precipitation
The ice erystal process

Since 1933 when BERGERON proposed [1] a mechanism for the formation of
rain by the initial development of snow in the upper parts of cool clouds,
it has been generally accepted that heavy rains could only be accounted for
in this manner. Experiences during the past war, especially in tropical regions,
Convinced fliers and some meteorologists that another mechanism could also
lead to the formation of heavy rain. Such precipitation was often observed
to fall from warm clouds.

BERGERON’s theory proposed that the difference in the saturation vapor
pressure with respect to ice and water would lead to the preferential growth
of an ice crystal at the expense of the cloud droplets in its vicinity. Without
doubt, this process is of major importance in the middle latitudes.

160 Vincent J. SCHAEFER ZAMP.

The vapor pressure differential process

Another precipitation mechanism parallel in some respects to that of
BERGERON has been proposed by PETTERSSEN [7]. This would account for the
development of rain drops in clouds due to the difference in temperature of
small rain drops falling into a warmer environment, the difference in saturation
vapor pressure of water at two temperatures producing a differential growth
in a manner akin to that of the ice crystal effect. A variation in temperature
of 0-01° C at 25° C would lead to the same differential as that which produces
the optimum condition for the growth of snow crystals.

While this mechanism would lead to a logical development of rain drops,
it is not easy to explain the method which starts the process. From a theo-
retical consideration, it is difficult to account for the early increase in size
of the cloud particles from their relatively stable dimensions. Lancmurr’s [8]
calculations suggest that it is impossible to maintain the necessary temperature
difference between small particles long enough for some of them to grow large
enough to start falling away from the others. From the experimental studies
in our laboratory, high velocities and large temperature differentials fail to
demonstrate this growth mechanism, although it is extremely easy to illustrate
the ice crystal effect.

The salt nuclei process

It is quite possible that the mechanism of natural rain formation in warm
clouds is intimately related to the presence of certain hygroscopic nuclei in
the air.

Woopcock has obtained experimental evidence recently [3] that there are
considerable numbers of large salt particles present in air over the sea in the
trade wind area. Salt particles as large as 2x 10-11g were collected by him
at altitudes ranging from a few meters to a kilometer above the sea surface.

A recent series of observations [9] by our Project Cirrus Group in the vicinity
of Puerto Rico showed that rain developed in extremely thin clouds below
the trade wind inversion at temperatures of +8° C. Light rains were measured
coming from clouds whose measured thickness was less than 300 feet. Fig. 5
shows a photograph of precipitating warm clouds. An estimate of the rainfall
rate calculated from the observed rate of collection in flight showed that it
would be approximately 0-05"/h. A similar measurement of a thicker cloud
2000 m (6500 ft.) in vertical thickness showed a rain water content falling
from the cloud into the sea at the rate of approximately 25 mm/h (1"/h). The
coldest part of this warm cloud had a temperature of +8° C.

Such amounts of rain from relatively thin clouds are rarely observed over
continental America and obviously require a mechanism for rain formation
different than the ice crystal effect.

Vol. I, 1950 Experimental Meteorology 161

While studying clouds in Puerto Rico, we also made some rough obser-
vations on the concentration of condensation nuclei in the air. Invariably,
with air coming in from the sea, the levels were extremely low, the number
rarely exceeding 200/cm%. Similar observations have since been made off the
New England Coast. In this latter case, however, the concentration jumps to
high levels within a few miles of the coast as the high levels inherent to con-
tinental air raise the concentration to between 10% and 105/cm*.

The experimental evidence thus far known seems to be compatible with
the view that the salt crystal nuclei are extremely effective centers for rain
formation. Due to their hygroscopic nature, they have considerable mass even
before a normal cloud forms. Due to the larger size of these particles, there
will be a tendency for them to start falling within the cloud at a different
rate than the small droplets in their vicinity, thus gathering in by collision
more and more of them. In this way, rain would form quite readily, and if
the cloud had a vertical thickness of a kilometer or so, the particles could
easily reach the maximum size that may fall without breakup. In addition
to the growth by collision as the drop became larger, it might also add to its
growth by the vapor pressure differential which, in the early stage, would be
somewhat enhanced because of its salt content and in the later stages by the
difference in temperature as the colder drop fell into warmer rising air.

Man’s efforts to produce rain

Down through the ages, man in various ways has tried to get rain, prevent
hail, and eliminate lightning. These efforts have ranged in method from cere-
monial sacrifices to rain gods to spraying electrified sand and dumping such
things as liquid air and dry ice to ‘“‘cool the air”. Since 1875, the literature
is replete with pamphlets, books, patents, and popular stories giving reasons
why such methods should be successful or detailed arguments to show that
they could not possibly be effective.

During the past thirty years, several methods have been tried to modify
atmospheric conditions which have claimed to make use of scientific prin-
ciples. Most prominent of these have been the activities of BANcRorT and
WarREN using electrified sand and VERAART using dry ice.

The electrified sand was used with the hope that the charged particles
would attract oppositely charged cloud droplets and thus form rain drops.
While considerable interest was displayed in these activities, the plan was
doomed to failure because of the tremendous quantities of materials required,
and the relative unimportance of the results obtained.

The next experiments which attracted considerable attention were centered
on the claims of VeRAART [10] of Holland who dropped dry ice from an air-
plane to affect clouds. A study of his theories and practices suggests that
162 Vincent J. SCHAEFER zame

he was merely putting into practice methods proposed by GATHMANN [11]
in 1891.

The effects expected from the introduction of dry ice into the atmosphere
was to reduce the temperature of the air to either form clouds where no clouds
previously existed or to augment the amount of condensed water already
existing as a cloud with the expectation that it would produce precipitation
of the cloud.

In his use of dry ice, VERAART proposed using very large quantities of this
material, He apparently failed to recognize the importance of the effect of
dry ice in supercooled clouds and, consequently, missed the chance to use
this material in an effective way.

Tn 1938 FINDEISEN [12] in concluding an important paper entitled, Colloidal
Meteorological Processes in the Formation of Atmospheric Precipitation, made
the following prophetic statement:

“The recognition of the fact that quite minute, quantitatively inappreciable
elements, are the actual cause setting into operation weather phenomena of
the highest magnitude, gives the certainty that, in time, human science will
be enabled to effect an artificial control on the course of meteorological phe-
nomena. It would be going beyond the limits of the present work to discuss
in detail the po j ng a kind of technical control over the course
of weather conditions. From the considerations under survey here, we have
now come to quite new points of view on this. It can be boldly stated that,
at comparatively moderate expense, it will, in time, be possible to bring about
rain by scientific means, to obviate the danger of icing, and to prevent the
formation of hailstorms. Through the energy transformations thus secured,
various other weather phenomena (e.g. temperature, wind) will be brought
under a certain kind of control, which perhaps never, in a direct manner,
could, to an appreciable extent, be acted upon in the atmosphere. The colloido-
meteorological investigations, by themselves with the only assistance of research
work on the means to get some control over the weather factors, have opened
up a new field for their efforts. They obviously only can solve those various
problems with the close assistance of aerology.”

Tn recognizing the possibility of modifying unstable cloud systems, Fixp-
EISEN pointed out the tremendous energies that might be released when the
proper type of ‘seeding agent" was discovered and properly used.

It is now believed that methods are now available to profoundly modify
cloud systems and thus realize some of the effects predicted by FINDEISE!

‘A study of the literature shows that at least one person observed that ice
crystals could be produced in air supersaturated with respect to ice if the air
was locally cooled to a very low temperature [13]. The significance of this ob-
servation as it might be related to meteorology was apparently not considered
by ADAMS.

Vol. 1, 1950 Experimental Meteorology 163

Meteorological studies at the General Electric Research Laboratory

In 1946, after spending the previous three years studying the nature of
snow storms and aircraft icing, the writer [14] described some laboratory experi-
ments concerning the seeding of supercooled clouds with dry ice. He pointed
out the important relationship between this effect, and the modification of
supercooled clouds in the natural atmosphere.

On November 13, 1946, the first experiment with seeding supercooled
clouds in the atmosphere was accomplished, producing results which had been
anticipated on the basis of the laboratory results. A four mile cold cloud
was profoundly modified within a few minutes by dispensing 6 pounds of
crushed dry ice into it. The cloud, which was supercooled to a temperature
of approximately —17-5°€ and which before seeding showed no sign of ice
crystals, was completely changed to snow within five minutes after seeding.

Subsequent experiments in the fall and winter [15] of 1946 included the
initiation of a snow area in the Hudson and Champlain Valleys of New York
the modification of a supercooled valley fog which initially reduced the visi-
bility to a remarkable degree, the production of snow showers from cold strato-

cumulus, and the production of extensive grooves and holes in a solid deck
of stratus clouds.

In addition to these flight experiments in the natural atmosphere, the
laboratory studies which had started in 1943 in the General Electric Research
Laboratory were carried out on an increased scale during the fall and winter
of 1946-7, :

These experiments
require extensive faci

pointed the way for further scientific work which would
he General Electric Company is not in a position
to supply such facilities and, consequently, a contract was initiated with the
Army Signal Corps. This contract is a joint Army, Navy, and Air Force ins-
trument wherein the General Electric Company provides scientific and tech-
nical guidance as consultants and the Government carries out all experiments
other than those done within the General Electric Research Laboratory. This
activity is identified as “Project Cirrus” and is administered by a Technical
Steering Committee consisting of representatives of the Army, Navy, and Air
For ©. Dr. IRVING LaycMurr and myself act as scientific consultants to this
ommittee,

At the present time, the Operations Group of Project Cirrus uses a flight
facility at the General Electric Flight Test Hangar at the Schenectady County
Airport which includes two B-17s, one PB4Y-1, one JRB, and one L-5 and the
necessary pilots, mechanics, cameramen, aerologists, and technicians to carry

out flight operations for studying the various phases of the precipitation
process.

164 Vincent J. SCHAEFER ZAMP-

The formation of a supercooled cloud

It is a simple matter to form a supercooled cloud [16]. A chamber having
dimensions of approximately 30cm wide, 50cm long, and 40 cm deep is
quite suitable for cloud experiments. Means should be provided to cool the
air in the chamber down to a temperature of at least —25° C, if possible. The
walls of the chamber should be painted black or lined on sides and bottom
with black cloth, such as velvet. Illumination may consist of a flashlight or
similar type of focused light beam. With the chamber!) cooled below ambient
room temperature, a cloud may be formed within it by the introduction of
moist air. Within a few seconds after condensation occurs, the droplets reach
the temperature of the chamber. Under ordinary laboratory conditions, the
cloud droplets reach a diameter in the range of 10 to 25 zm and a concentration
of 200 to 1000 per cubic centimeter. Occasionally, a few ice crystals appear in
the chamber if the temperature is colder than —10° C. Generally, however,
this is a transient effect with rarely more than 1 crystal/cm® forming. The
supercooled cloud droplets persist until the air is no longer supersaturated
with respect to water. During this period, a wire or miniature propeller will
be coated with ice if rotated within the cloud. Eventually, it disappears, the
droplets slowly settling to the bottom of the chamber or evaporating onto
the frosty walls.

Such relatively stable supercooled clouds can be formed to a temperature
of nearly —39°C. If, however, the temperature is reduced below this value,
it is impossible to form a supercooled cloud!

The formation of condensation nuclei

In most air samples likely to be used in laboratory experiments, there is
no lack of condensation nuclei. Concentration ranging between a thousand
and a million per cubic centimeter are normally observed. If the level is low,
it may be increased by burning a bit of charcoal, striking a match, heating
a nichrome filament, sparking a Tesla coil, or atomizing a salt solution. In
fact, some very striking experiments may be carried out to demonstrate the
optical effects possible with variations of type and concentration of conden-
sation nuclei.

The formation of sublimation nuclei

To demonstrate the presence of natural sublimation nuclei in the air under
laboratory conditions is not easy. Sometimes the free air contains relatively

1) A very convenient type of chamber is a 4 cubic foot home freezer, although it is possible
to conduct effective experiments with much cruder apparatus if necessary. Two galvanized tubs
separated by a water-ice-salt solution may be quite adequate for short experiments.

Vol. 1, 1950 Experimental Meteorology 165
high concentration with the number occasionally reaching 10 particles per
cubic centimeter. However, under ordinary conditions, the concentration seems
to range between 50 and 500 per cubic meter. Under such conditions, a few
particles will be seen in the beam of a light directed into the cold chamber
containing a supercooled cloud.

As indicated in an earlier chapter, certain clays and other mineral dusts
serve as effective sublimation nuclei at definite temperature ranges below 0°C.
A given sample my be evaluated by using particles of such a size that they
readily form an aerosol. Shaking a box containing the sample while held in
the chamber will produce a cloud, since the fine particles will float out of
the container. The particles, if effective as ice nuclei, will become coated with
a frost layer if the air is supersaturated with respect to ice. This normally occurs
within 30s, They become visible as twinkling crystals if allowed to grow and
generally form asymmetrical crystals unless the initial particles are smaller
than 1 yam in diameter, Methods already described in detail [16] may be used
to study such particles.

The role played by silver iodide serving as a sublimation nucleus is out-
standing. It may be introduced into the air some distance away from the
laboratory and still have an appreciable effect in the laboratory if the air
trajectory is favorable. A few simple laboratory experiments will be mentioned.
\ wire filament dusted with a few minute particles of silver iodide will, if heated
in air supersaturated with respect to ice, produce many millions of ice crystals.
The particles formed in this way are submicroscopic with many less than
100 A in diameter, By drawing an arc with a pure silver wire using either
a Tesla coil or by momentarily shorting the leads of a dry battery with a
silver wire, a smoke of silver particles may be introduced into the cold chamber.
In a supercooled cloud formed subsequently, no ice crystals will be observed
if iodine vapor is absent. If, however, a small iodine crystal is then passed
briefly through the chamber, large numbers of ice crystals will be seen to form
in the wake of the crystal within the supercooled cloud. A still simpler means
of demonstrating the silver iodide effect is to place a few particles on the end
of a match which will be volatilized as the match is ignited, thus producing
many nuclei.

As shown [5] by VonNEGuT, the nature of silver iodide smokes in forming
sublimation nuclei seems to be related to a probability function which has a
fairly high temperature coefficient. Thus, at a temperature of —6°C, some
sublimation nuclei will appear. At —10°C with all other factors the same,
many more particles will be observed in the same unit time. Space does not
permit a detailed discussion of the interesting relationships which have been
found, The original reports and papers of VONNEGUT should be studied care-
fully if plans are contemplated to use this material for laboratory or field
experiments.

166 Vincent J. SCHAEFER ZAMP.

Establishment of Project Cirrus

Early in 1947 under the direction of Dr. Irvinc Lancuuir assisted by
the writer, a much more extensive cloud studies program was initiated by the
Research Laboratory under the sponsorship of the Army Signal Corps, the
Office of Naval Research, and the Air Forces. Under this arrangement, a tem-
porary flight facility was established at the General Electric Flight Test Hangar
at the Schenectady County Airport which, at the present time (year of 1949-50),
includes 2 B-17s, 1 PB4Y-1, 1 JRB, and 1 L-5 and the necessary pilots,
mechanics, cameramen, aerologists, and technicians to carry out flight oper-
ations for studying all phases of the precipitation process.

Project Cirrus and experimental meteorology

Project Cirrus is a fundamental research study of the physical and chemical
processes which occur in the lower atmosphere and produce clouds, snow, rain,
atmospheric electricity, and associated phenomena.

One of the major activities of Project Cirrus is in the field of experimental
meteorology. In this respect, observations and experimental studies are made
with cloud systems having volumes ranging from less than one to more than
several hundred cubic miles. The limitations in size are related primarily to
available cloud systems and the physical conditions required for each partic-
ular experiment.

These studies are planned and carried out insofar as possible as laboratory
type experiments. Each operation is planned to supplement others previously
accomplished so that some features are checked as new aspects are under
exploration. Insofar as possible, controls are maintained so that comparative
evaluations may be achieved.

The major objective in this particular phase of our study has been directed
toward the detection of unstable atmospheric conditions which develop in the
atmosphere and often persist for some time. When such conditions are dis-
covered various “triggering” actions are then applied in an attempt to shift
the system to its more stable form. Such modifications of clouds often involve
the release of tremendous quantities of energy. As pointed out recently by
LonGLeyY [17], the energy release following the condensation and subsequent fall
of one inch of rain on one square mile of the earth is equivalent to 1-7 x 10! erg.
For comparison purposes, this is about twice that of the energy said to be
released by an atom bomb of the type dropped on Hiroshima. Although much
of this energy is released as the cloud forms, unless it is precipitated on the
earth in an effective and useful way, it may be regarded as lost energy insofar
as the earth and its water resources are concerned.

Evidence is now accumulating which shows that under certain conditions
it is feasible to initiate the precipitation cycle artificially in some types of cloud

Vol. 1, 1950 Experimental Meteorology 167

systems so that their increased output forms a valuable addition to the natural
resources of the earth.

Before going into a discussion of this fascinating subject, it may be in order
to review briefly the operational facilities and procedures which we are now
using in our basic research and studies of clouds and the atmosphere.

These acti may be divided into three major parts: (1) laboratory
research; (2) field studies; (3) flight operations. 7

Laboratory research under Project Cirrus

The main laboratory studies are conducted at the new General Electric
Research Laboratory building at the Knolls in the lower Mohawk Valley in
eastern New York. Complete facilities for physical and chemical studies with
many unique features are available in this laboratory. Part of the laboratory
in use includes a weather observatory equipped with standard, as well as
special, meteorological instruments, radio communications, a small wind tun-
nel, and a complete photographic dark room. In addition to the laboratory
areas, excellent shop facilities are available including the services of skilled

technicians for special developments in mechanical and electronic instrument-
ition.

In addition to these facilities, meteorological observations of a specific
nature are carried on by special observers at the Mt. Washington Observatory
on the summit of Mt. Washington in the state of New Hampshire. This moun-
tain observatory, at an elevation of 6288 feet above sea level, is world famous
for its exceptionally severe weather, especially high winds and extended rime
storms produced by supercooled clouds sweeping over the mountain. Besides
having projects carried out by Observatory personnel for Project Cirrus, the
research facilities on the mountain are always available to members of Project
Ginna for testing instruments and studying various types of natural orographic
clouds.

In addition to the experimental research activities at the Research Labo-
hitory, an important phase of the laboratory program is the analysis of flight
data by photogrammetric methods and detailed studies of meteorological
conditions present during the experiment. Considerable space in the laboratory

is used for these purposes, and at least one person spends full time on this
activity,

Field studies under Project Cirrus

Supplementing experimental research in the laboratory, the Research
Group is involved in a considerable variety of field studies ranging from a
study of the effect produced in the atmosphere by ground generators dispens-
ing silver iodide smokes, detailed observations of various types of natural
168 Vincent J. ScHAEFER AMP

rain and snow storms, studies of the development of all types of clouds using
lapse time motion pictures, determination of the concentration of condensation
and sublimation nuclei which occurs in the atmosphere, and activities of a
similar nature dealing with weather phenomena associated with the formation
of clouds and the subsequent development of precipitation. Many observations
are made in regions other than eastern New York. For example, cloud studies
have been conducted by one or more members of the Research Group in
northern Idaho, Wyoming, and other parts of the Northwest, Florida, Puerto
Rico, various parts of New England, and Central America, particularly in
Honduras. It is of great importance that cloud systems in various parts of
the world be studied and their local peculiarities understood, since it is quite
obvious that large differences exist in clouds, not only in their general develop-
ment and life cycles but even in their microstructure. Until these variations
are better understood, it will be impossible to draw general conclusions about
them.

Flight operations by Project Cirrus

At the present state of our knowledge of clouds, it is of great importance
that the general structure, as well as the microstructure, of clouds be explored
by going into them. This may be accomplished to a limited degree by observing
them at a mountain observatory using the summit as a stationary probe to
study their structure as they pass by the station. The information gained in
this manner is of great value but does.not provide much data on convection,
turbulence, and three dimensional structure which is of basic importance in
studies of clouds in the free atmosphere.

The only method now reasonably satisfactory involves the use of one or
more aircraft which can probe clouds at various levels and in doing so, register
on automatic instruments some of the properties characteristic of the clouds
explored. Good photographic techniques are of extreme value in this respect
because of the complexity of clouds and the rapidity with which they change
some of their features. It is impossible to obtain a satisfactory record by
visual observation alone.

In this respect, some of our seeding techniques are of unique value since,
for the first time, they provide a method of marking a cloud that will persist
for a long time and may be seen for large distances. By taking consecutive
pictures of a cloud area marked in this way, much information may be obtained
during periods of an hour or more which shows the various mechanisms that
are of importance in the formation and dissipation of clouds.

To carry out such a flight program effectively requires much organization
and specialized training of the personnel engaged in the work. The planes
involved must have a considerable amount of workable scientific equipment
especially suited for meteorological studies. In addition, it is of great import-

Vol. 1, 1950 Experimental Meteorology 169

ance that a special schedule is followed in reporting the results of each ex-
perimental flight study.

A typical flight operation of Project Cirrus

It may be of interest to describe a typical experimental flight study con-
ducted by the Project Cirrus Flight Operations Group. For this example
Flight 83 will be described since it was a two plane operation employing both
dry ice and silver iodide in the seeding operation. :

At 1500 on the previous day, the weather group assigned to Project Cirrus,
comprised of Navy personnel, notified the Chairman of the Operations Group
that the synoptic situation suggested the strong possibility that a suitable
supercooled deck of stratus clouds might be expected the next morning. An
alert was sounded, crews were assigned duties, and tentative plans set for
0900 take off in the morning using two B-17 planes.

Early the next morning, a check on the weather developments showed
that the forecast was good, and each member of the flight group carried
through his assigned duties prior to take off. These duties included, besides
normal preflight preparations, such extra things as cleaning the windows used
by the photographers, crushing and packing the dry ice in canvas bags, load-
ing the silver iodide dispenser with fragments of impregnated charcoal, load-
ing and checking the photopanel camera for operation, checking the operation
of the automatic recording instruments, and making sure that the inking
pens and charts were ready for operation.

Except for the preparation of the silver iodide and dry ice and their dis-
pensing mechanisms, the above special activities were required for both planes.

On B-17 No. 5667 used as the seeding and probing plane, a crew of ten
men were assigned for the operation including:

1 Pilot
1 Co-pilot
1 Navigator
1 Flight Controller (alternate)
1 Photographer
1 Technical Observer
1 Aerologist
1 Radio Operator
2 Flight Mechanics
On B-17 No. 7746, the photographic and observation plane, a crew of seven
men included the following:
1 Pilot
1 Co-pilot
1 Flight Controller

170 Vincent J. SCHAEFER zaMe

1 Photographer

1 Aerologist

1 Radio Operator
1 Flight Mechanic

Just before take off, both flight crews assembled for a briefing at which
time a brief description of the proposed flight was given by the Flight Con-
troller, including the general objectives of the operation. After take off, the
planes were to rendezvous over the Albany Radio Range at approximately
20,000 feet. When the rendezvous was accomplished, the planes would again
check radio contacts and proceed together to a position about 30 miles NW
of the range with the seeding plane holding a position on top of the stratus
deck, the photo plane climbing but maintaining visual contact with the lower
plane.

From his vantage point in the photo plane, the Flight Controller sized up
the situation as favorable for a “Figure Four” seeding pattern using */, pound
of dry ice pellets per mile of flight with a short seeding with silver iodide. He
then ordered the preparations to start for dispensing the silver iodide char-
coal since the seeding flight would start within a few minutes. The proposed
“Figure Four” flight plan as suggested by the Research Group and adopted
by the Operations Group is shown in Fig. 6.

‘As soon as the order to seed was given, the seeding plane went into its
pattern, flying several hundred feet above the top of the slightly ragged top
of the stratus deck putting out dry ice pellets for five miles, after which for
one mile no seeding agent was used. The order to dispense the small burning
charcoal fragments producing silver iodide was then given. About 20 seconds
was required to dispense the silver iodide particles. Another one mile gap
without seeding was next ordered, after which time dry ice seeding was again
started and continued at the rate of 1% pound per mile until the “Figure
Four” was outlined. Following this, a single dry ice pellet drop was made
bracketed by a line of continuous seeding parallel to the first leg of the four
pattern. 7

Throughout all of the seeding operation, the photo plane was cruising at
21,300 feet. For the first time, the seeded track immediately behind the seed-
ing plane was photographed. Fig. 7 illustrates one of several remarkable photo-
graphs obtained at this time showing the speed with which the dry ice effect
spreads in the wake of the seeding plane.

Shortly after take off the photopanels in both planes were started. Each
panel holds the following instruments:

Rate of climb
Air speed
Altimeter
Bank and turn

Vol. 1, 1950 Experimental Meteorology 171

Compass

Clock

Counter

Battery of station indicator lights

Automatic photographs are obtained of this instrument panel every
55 seconds while, in addition, whenever any one of four other switches at
various positions are tripped an additional picture is taken. This permits the
photographers, the aerologists, the flight controller, special observers or those
with special assignments, such as dispensing the seeding agent, to produce
4 special record of any operation he might make individually as part of the
flight operation,

After the seeding flight was completed, plane No. 5667 was directed by
the Flight Controller to obtain low level photographs of the developing seeded
pattern and, in addition, to probe the infected area to observe optical pheno-
mena and other features that might be of interest. Fig. 8 illustrates typical
photographs obtained from this plane. Since the flight was made over moun-
(ainous terrain, it was decided to forego a descent through the deep trough
cut by the ice crystals. After 30 minutes of probing studies and low level
photographic coverage, it was released from further cooperative observations
by the Flight Controller. A total of 24 photographs were taken at various
altitudes up to 3000 feet above the cloud deck.

Meanwhile, plane No. 7746 was cruising above the seeded track taking
till photographs and a few moving pictures. This continued for a period of
{0 minutes during which time a total of 48 photographs were obtained. Figs.
and 10 are typical photographs taken during this period. By this time, the
‘light Controller decided that an adequate set of photographs had been ob-
tained since the pattern was beginning to deteriorate and no new phenomena
were in evidence. It should be mentioned at this point that throughout the
observation flight, the entire crew in plane No. 7746 were on oxygen since
the flight occurred at an altitude of 21,300 feet. The flight was then terminated
with both planes heading for the base where they landed at about 1130. Thus
the operation required a total flight time of approximately 150 minutes,
about 40 percent of this time being employed in the experimental studies.

Procedure for reporting on an experimental jlight study

lhe procedure now in use by the Flight Operations Group to effect a close
relationship with the Research Group may be divided into four stages.

The preflight briefing - This involves the development by the Research
Group of a series of experiments required to provide specific data on certain
types of clouds. These requirements are studied by the Operations Group and
adopted to flight procedures. A Flight Controller is designated who is directly
172 Vincenr J. ScHarrer ZAMe

responsible for the carrying out of the complete experimental flight. Personnel
are kept on the alert for suitable cloud systems whenever the aerologists report
suitable conditions within 200 miles. As soon as reasonable assurance is at
hand that the type of clouds needed may be found, the planes take off, approach
the system, and then follow through according to the briefing plan.

It is of great importance that the Flight Controller has several alternate
plans for immediate substitution in the event that the cloud system is some-
what different than expected from the reported synoptic situation. He must
have the ability to size up the situation while approaching the experimental
area and thus take advantage of whatever cloud system is found.

The preliminary report - Within an hour after the flight is completed,
a brief report is transmitted to the Research Group by the Flight Controller
which summarizes the results as observed and the general data obtained. This
includes enough detail so that it is possible for the Research Group to deter-
mine immediately whether there is need for another flight to supplement the
data obtained.

Detailed flight report — A more detailed report including copies of all the
raw data, all logs and notes of observers, and contact prints of still photo-
graphs obtained are supplied to the Research Group within two days. Follow-
ing a review of the contact prints, the Research Group orders enlargements
of all photographs which appear suitable for analysis.

Final report — Within a week or two, all supplemental data not available
at the time of the second report, reduced data from the photopanel, enlarge-
ments of selected prints, moving picture film and a meteorological analysis
of the weather preceding, during, and subsequent to the flight study are sup-
plied to the Research Group to aid in the analysis. The detailed analysis of
the flight is then scheduled with relation to other flight operations under
study. The summary of this work is subsequently published in an Occasional
Report of Project Cirrus.

It should be pointed out at this time that the above procedure is carried
out beyond the third and fourth stages on only those operations where the
accumulated data is reliable and has sufficient detail to warrant spending
the time involved in carrying through a complete analytical study.

Types of cloud seedings used in Project Cirrus operations

Two types of cloud seeding as related to flight studies shall be discussed
at this time—the formation of ice crystals in supercooled clouds and the
development of a chain reaction in cumulus clouds using large water drops.

The dry ice effect - When solid carbon dioxide is introduced into a super-
cooled cloud in the atmosphere, enormous quantities of ice crystals form and
produce profound changes in the cloud by the mechanism explained by the

Vol. 1, 19

Experimental Meteorology 173

Ie RGERON-FINDEISEN theory. Crystals may be produced in such large quan-
tities that it is sometimes possible to create conditions unlike those which
occur naturally in the free atmosphere.

As mentioned earlier, it is extremely uncommon to find any ice crystals
forming in natural clouds until some region has a temperature of —12°C.
Since the introduction of carbon dioxide ice (henceforth called dry ice) will
produce ice crystals at any temperature below 0° C, many cloud systems that
would not shift to snow naturally may be affected by this type of artificial
inoculation.

The enormous numbers of snow crystals produced in this manner also
illow us to modify unstable supercooled clouds in several ways which rarely, if
ever, happen in nature. For example, it is possible by artificial means to shift
all of the condensed water in a massive supercooled towering cumulus cloud to
now crystals in considerably less than five minutes. A similar cloud by natural
processes normally requires an hour or more to reach the same condition and
even after that time, might not be completely modified. The quantity of dry
ice required to accomplish artificial modification is insignificant, since, as
pointed out in previous papers [14], one gram of dry ice is capable of pro-
ducing at least 1018 ice crystals. Fig. 11 is a photograph showing the cloud
als that trail from a single piece of dry ice falling through air.
Laboratory experiments show that tremendous numbers of ice crystals stream
from a smal] pellet as it falls through the air. In order that these crystals
become effective, they must form in air colder than 0° C which is supersaturated
with respect to ice. By flying above or through the cloud, dry ice particles
having sizes ranging from 1 to 20 mm in diameter fall down or are carried
loft, depending on their size and the turbulence in the clouds. Natural
convection and turbulence of cumulus clouds, augmented by the heat released
by the change in phase from water to ice, assist in causing the rapid infection
of a large region of the cloud system. Experiments show that if a concentration
of crystals exceeding about 50 per cubic centimeter is present, the supercooled
cloud thus infected is completely evaporated in less than 10s. With a
concentration ten to twenty times more than this, the competition between
particles for the available cloud water is so great that none of the particles
yrow as large as the original supercooled cloud droplets. As a result, the cloud
iy ‘overseeded”’ and becomes extremely stable. Such overseeded clouds rapidly
Joye their convective activity and become very stable and persistent. Examples
of overseeded clouds may often be seen to form naturally when cumulus clouds
pass through the transition temperature of —30° C. Fig. 12 illustrates a typical
cloud of this sort. This generally occurs at altitudes of 28,000 feet to 32,000
feet, the results appearing as anvil tops or as long streamers of cirrus clouds
drifting across the sky in the tropopause region of the atmosphere. These
overseeded effects, however, are rarely, if ever, observed in the natural atmo-

of ice crys

Vol. 1, 1950 Experimental Meteorology 175

174 Vr

ent J. SCHAEFER ZAMP.

pressure grow by diffusion since they are continually invading air that is
warmer than the residual temperature of the falling drops. When the drops
reach a weight of about 0-5 g, they are no longer spherical but are flattened
out in the peculiar manner shown in Fig. 13. Such drops are potentially
unstable and become susceptible to break-up. A small shearing force of the
kind common to turbulent air is all that is required to shatter the drops into
two or more smaller drops. BLANCHARD’s experimental studies [18] in our
laboratory show in a very elegant manner the limiting conditions of stability
which restricts the size of falling rain drops.

If a growing rain drop in a cumulus cloud breaks apart before reaching
the level from which it started growing, the mechanism constitutes a chain
reaction. Thus one drop produces two or more; these droplets passing through
the same cycle produce two or more, and within a very short time, many
uillions of particles have developed from the initial drop. If, for example,
tich an unstable drop breaks into five droplets (a common occurrence as
wen in BLANCHARD's experiments), it requires only 10 cycles for more than
2 million new drops to form.

The effects which may develop from water seeding of cumulus clouds — Under
inost conditions, the introduction of large water drops into a convective cloud

sphere at warmer temperatures unless the clouds are seeded artificially or at
times when natural seeding takes place due to the entrainment of snow crystals
carried down from higher altitudes through stable air.

Since it is demonstrable that cold clouds may be overseeded, it follows
that lesser amounts may be introduced as desired. This makes it possible to
produce many interesting effects in cold and cool clouds. For example, it is
feasible to seed and dissipate clouds at specific altitudes above the freezing
level in the atmosphere and thus prevent the development of large vertical
thicknesses of supercooling. In this way, thick supercooled clouds cannot form.
Consequently, the sudden release of a large amount of energy necessary to
produce thunderstorms and similar disturbances may be checked or at least
reduced in intensity.

On the other hand, if it is desirable under certain conditions to go to the
other extreme and attempt the release of the maximum amount of energy
possible from a particular cloud system, this may be accomplished also. For
this to be successful, it is necessary to wait until the maximum vertical de-
velopment occurs at which time the cloud is seeded in such a manner that the
optimum number of crystals are introduced to cause a rapid shift from the
water to the ice phase and, at the same time, obtain the most effective particle

. ere a coat . 3 exceeding these critical dimensions might be expected to do no more than
size to initiate rapid precipitation. By properly carrying out such operations, i ivi4e a chain reaction within the cloud which would lead to its dissipation.
itamight at times (be-possible to félease-encugit snery to breaie through: an Since this mechanism is a mechanical one and is not related to a change
versions limiting the vertical development of the clouds. i, phase, no energy release is effected and, consequently, one would not nor-

The development of Precipitaion by water seeding — LANGMuTR has pro-
posed [8] a mechanism for initiating precipitation in cumulus clouds. The method
involves the introduction of relatively large water drops into clouds having
the following properties:

(1) They must be actively growing cumulus clouds having vertical thick~
nesses greater than 1-14 km.

(2) The upward vertical convection in some region of the clouds must,
exceed 2-14 m/s.

(3) The droplets in the cloud must have a diameter of 15 ym or more.

(4) The average liquid water content of the clouds should exceed 2-7 g/m*.

Field studies show that most actively growing cumulus clouds having a’
vertical thickness of 1-14 km possess the other characteristics mentioned above.

When water drops larger than 0-01 cm diameter are introduced into
region of the cloud having strong upward convection, they sweep up the smaller
droplets in their path as they are carried aloft. Most of this coalescence occur:
on the under-surface of the drop since the cloud droplets move faster than.
the larger droplets and are thus intercepted and collected. When the growin

inully expect anything to happen, save the disappearance of the cloud.

If, however, the precipitation develops as a chain reaction so that heavy
tain forms on one side of a large convective cell, the down drafts might be
© strong that an upward counter current is produced. If the air is unstable,
this upsurge of air may lead to a local convergence which would certainly
produce more precipitation than would normally result by the mere dissipation
of the treated cloud.

The normal effect, however, which seems to be most commonly experienced
when clouds are seeded by water is dissipation. This is an important feature,
however, since even supercooled clouds may be affected in this manner. In
fespect to the dissipation of clouds, it should be emphasized at this point
that many large clouds dissipate naturally, especially when the air aloft is
dry, The best way to evaluate such results is to become familiar with the
yrowth and disappearance of clouds by the use of lapse time moving pictures.
Successive pictures of clouds taken by a movie camera at 2-14 second inter-
vals and then viewed at the normal rate of 16 per second, speed the apparent

Pee onede the vectical Iift of thelp l0\elopment of clouds by a factor of 40 fold. A familiarity of cloud develop-
drops become so large that the pull of gravity exceeds the vertical lift of theg ii) ined in this manner permits the observer to make a critical evaluation
air current, the drops begin to fall and sweep up even more cloud particles ii which follow seeding operations.

in their path. In addition, the falling drops by virtue of their lower vapoi
176 Vincent J. SCHAEFER Zane

Volt, 1950 Experimental Meteorology 177
Table 1
0%
Mass of water vapor in saturated air (from Smithsonian Tables)
Temperature | Amount of water vapor
° in saturated air g/m* tore
30-039
17-118 ovo
9-330
4835
2-154
0-892 tore
WARM CLOUD COOL CLOUD
orc
ae _ _——
Table 2

Concentration of ice nuclei in air of Northeastern U.S.
(January 1948 to September 1949 [3 hourly observations})

Nie 1. ‘Temperature relationships in warm, cool, and cold clouds.

Number of ice nuclei_ | Number of observations
per cubic meter of air | during 18 months period

1X 10° to 1 x 10? 1194
1x 10? tol x 104 1294
1x 104 to5 x 105 1757
5X 105 tol x 107 295
ee -30°% Ly75GM*
Ate
ih fy ”
. -3 _
Lg O°C Ly89GM — 10°C Ly 05 GM?
Oe
f 295°C CLOUD BASE 4 ee eee
Ter EVEL
Viet Approximat a: m values of the liquid wa cor supercooled 1 IS a
(Received: October 1, 49.) (To be continued in ZAMP 1/4) Ma ivaiyys clomle at eoveral soon fa Oe Me eeatent ee vaeea

Wane tie
78 Vincent J. SCHARFER ap 160 Experimental Meteorology 179

-40 -30 -20 -0 ¢ oO

(CARBON DIOMIDE TNE GADEN lan vsoe provucts pian efrect. J)
SILVER IODIDE —
LOAM,RUGBY N.D.
(CLAY, GUILDERLAND NY.
LOESS|HANFORD WASH.
'LOAM, BRUEL|WIS.
SOIL ,WOLF PT. MONT.

SOIL,BAGGS WYO. oe
SAND, AGATE (colo. |

‘LOAM, GOURD'ALENE_10.—>—
ASH,CRATERLAKE ORES

KYANITE Al2 Si0s

eee CUTIN MEX. 1 |
[DUST,PHOENIX ARIZ. —
MARL,RAVENNAINY. |= — |
[BENTQNITE, MONT. ——
SOIL,NEV. [+
DIATOM!
SPOR) |
LOAM,OAKLEY KANS. EE,
KAOLIN\GEORGIA

1s sublimation
of certain

nay form

a

INDICATES THRESHOLD OF
_——*8 ACTIVITY
INDICAT! COMPLETE ACTIVITY

180 Vincent J. So zAMP Vol ty tase

Experimental Meteorology 181

Vie 7, View of rapidly spreading seeding track behind plane dispensing dry ice. Seeding plane several

Nundred feet above cloud deck. (Official Photo, Signal Corps Engineering Laboratory.)
Fig. 5. Rain showers from warm tropical cumulus cloud without appreciable v development

SINGLE PELLET DROP

10:19:48 0:19:11
4 10:18:28
10:20:02 co ie 4
* 05:37
lor2or2t y ees

~+— NOSEEDING TRACK
CO, SEEDING TRI
xx AgI SEEDING TRACI

}

“onanism

TRUE AIR PLOT SEEDING TRACK
We FLIGHT 83
ot23 45

NAUTICAL MILES

Fig. 6, Seeding pattern used in Flight 88, Project Cirrus.

ing area of snow crystals produced by dry ice seeding, Flight 73, Project Cirrus.
(Official Photo, Signal Corps Engineering Laboratory.)
Vincent J, SCHAWHNK Vol. 1, 1950 Experimental Mete

Fig. 9. View of “Figure four” seeding, Flight x3, Project Cirms (Oifelal Phate,
Engineering Laborator

ig. 10. View of “Figure four” pattern from near apes, HIME Ml HAMIL Hii Siynal Corp
Engineering Laboratory.) Fig. 11

Ice erystals streaming from pellet of carbon dioxide ice.

ND
a

Zam 1, Avon Experimental Meteorology

184 Vincent J. SCHAEFER

ig. 12. View of cumulus cloud generating large quantities of ice crystals. Streamers of “false cirrus"

produced by successive build-ups of cumulus.

Ihe application of seeding methods to clouds of various types

The modification of orographic clouds

clouds by seeding methods presents a parti-

Ihe modification of orographi
ularly intriguing possibility. These are clouds which form as moist air is forced
tor is it encounters a barrier such as a mountain range. In rising, the air
junds and cools as its pressure decreases. If the amount of cooling drops
(le uir below the dew point temperature, a cloud forms

\ider many conditions, the clouds formed by orographic processes are

or cold clouds, especially in the wintertime or when towering cumulus

}) over certain mountain peaks or ridges.
Ihe condensed water in orographic clouds in the wintertime is not very
liiyh, since it rarely exceeds 1 g/m. Such clouds, however, are very common

i} Mountainous regions and often form continuously for many days. Even a
timmory study of them reveals that relatively little precipitation reaches the

s or in the form

Wi from them except as rime deposits on trees and roc!
{ scattered snow crystals. Under most conditions observed on mountains in
(he hortheastern United States, snow crystals do not form in sufficient numbers
(i) tie up the available supercooled cloud droplets. Consequently, only a small
fraction of the clouds which form in this manner are precipitated. If techniques
unt ised to cause a widespread and effective precipitation of such clouds,
the depth of the snow pack in the vicinity of mountains might be markedly
iieveased, Such a result would be of much importance since the snow pack
streams which

Mi Mountain slopes is of great importance in stabilizing the

flow from such regions.

Fig. 13. The oscillations and breakup of a single large drop of water floating in air having a velocity
equivalent to the terminal velocity of the drop. Illumination with high intensity stroboscopic light

tric Research Laboratory Photo.)

round generators using silver iodide smokes is one way in which

“iagraphic clouds might be seeded. Unless such clouds form at relatively low

us

(General El

218 Vincent J, SCHAEFER ZAMP

temperatures, however, this seeding material will not be of much importance
since temperatures below —10°C are necessary if an efficient production o
nuclei is to occur. The concentration of nuclei must be of the order of 50 t
100 per cubic centimeter where the vertical rise of the cloud is rapid if th
available cloud droplets are to be converted to snow. Whether particles of this
concentration will subsequently grow large enough to form precipitation is

question which is answered best by experimentation with varying types o!
clouds, temperatures, wind velocities, and vertical accelerations in mountainou:
regions. Such particles will probably grow large enough to fall to the earth’
surface if the cloud beyond the mountain summit is of sufficient thickness t
sustain the growth of the particles until they become large enough to preci
pitate. In many instances when orographic clouds have temperatures of —10°

or colder, the liquid water content of the clouds is so low that it is questionabl
whether the precipitation initiated with silver iodide would be economicall
feasible.

The production of ice crystals in orographic clouds by the use of dry ice
liquid carbon dioxide, or similar methods requires that the crystals be intro
duced into air supersaturated with respect to ice. In addition to this require:
ment, it is also necessary that the seeding be a continuous operation. Thi
imposes rather severe limitations on the sites where such experiments ma:
be carried out effectively. For this reason, it is at present questionable whethe
a feasible method is known for carrying out the seeding of relatively thin an
cold orographic clouds on a scale that would have economic importance.

A method might possibly be developed, however, making use of the rimin
nature of cold clouds so that very weak rime feathers are formed which cot
tinually shed tiny ice fragments into the wind. That this might be feasibl
is suggested by the fact that small rime fragments form a considerable pet
centage of the ice particles observed in the air during an icing storm. Tf th
source of these particles was better understood, a more effective way of for
ing them in larger quantities could probably be developed which might no
require more than the planting of certain types of sub-alpine trees or th
construction of man-made structures in strategic regions on the upwind regio
of the mountains.

The production of regions of ice nuclei in the sky

It sometimes happens that large snow storms from low, cold clouds a1
started and kept going by their contact with a thin layer of middle or up
clouds, such as altostratus or cirrus. CONOVER has described [19] an interesti
case of this kind.

It is a fairly common experience to note examples of the seeding of low
clouds by cirrus crystals. This latter type of observation may be deduced by

Volt, 1980. Experimental Meteorology

219

\ study of the snow crystals reaching the ground during a snow storm. This
‘ondition generally produces stellar snow crystals with cirrus type hexagons
in the center. This is illustrated by Fig. 14 which shows a few samples of
cryatuls which grew in this manner.

Ihe production of specific regions in the free atmosphere containing high
concontrations of ice nuclei or potential ice nuclei is now an interesting possi-
hility, Cold middle clouds, even though having no appreciable moisture, may
lw sed as “holding reservoirs” to store ice crystals until they come into
contact with lower clouds of greater thickness or are entrained into cool or
‘old cumulus. An example of this type of seeding is contained in the seeding
operation during our high level study [20] of Hurricane King on October 13,
1017, \ relatively thin layer of stratus clouds covering an area of nearly 300
jwure miles was transformed to snow crystals. The subsequent fate of these
ive crystals is still a moot question, but if a considerable region of them was
entrained into the lower levels of a line of towering cumulus observed during
ihe flight located in the southeast quadrant of the storm, the entrainment
af (hese snow crystals might have exercised a profound effect on the subse-
quent development of these cumulus.

Similarly, the ice crystal residue from seeded, but small, cumulus clouds
thay be entrained at a low level into much larger cumulus forming in their

Vicinity, In this way, an effect of considerable magnitude is produced as the
up nepoled regions are infected at a lower level than would otherwise be
pownible

It will take much careful study to establish methods for utilizing this type
of svoding. Eventually, it may become one of the most important of all.

\ discovery that would have great importance in this respect would be a
(ible sublimation or freezing nucleus which would be effective at a tempe-
/iHire within one or two degrees below 0° C. It is obvious from the observations
jwle thus far that natural nuclei of this kind are rarely, if ever, formed.
Ii (hus remains for us to find or develop a substance in the laboratory which
will fit the requirements. :

The modification of stratiform clouds

Ihe widespread modification of stratus clouds by artificial means is pos-

‘ible at the present time whenever such clouds are supercooled. Under such
conditions, the clouds may be further stabilized by overseeding them or their
jeripitition may be accomplished by using an optimum number of ice nuclei.
Hhiis latter result is achieved by using only enough ice nuclei to cause the cloud
juirticles to evaporate completely as they condense onto the crystals thus formed

Which then grow large enough to fall as snow.

lypical results obtained in seeding cold stratus clouds are shown in Figs.
18, 16, 17, 18, 19, and 20.

220 Vincent J. SCHAEFER vamp

Stratus clouds may be seeded by flying 30-100 m above them and dropping
dry ice fragments ranging in size from 0-1 to 1 em diameter at the rate of
approximately 250 g (1% pound) per mile. Except with clouds thicker than
2 km (6500 ft.), the use of more than one pound of dry ice per mile will tend
to produce overseeding.

Besides seeding stratus clouds from on top, it is 4
effectively by flying through them as well as by flying at the cloud base.
At this lower position, however, there is no need of using large fragments
since the dry ice is only effective in air supersaturated with respect to ice
‘A zone below the cloud base equivalent in depth to approximately 10 m per
degree centigrade below freezing will support the formation of ice crystals
formed with dry ice. Thus, if the temperature at the cloud base is 10°C,
a distance of 100 m below the cloud will become filled with ice crystals if dry
ice fragments are sprinkled or liquid CO, is sprayed into that region

Aso possible to seed them

The modification of supercooled ground Joys

While warm ground fogs formed by advection or radiation may only be
modified at present by heating the air to cause its evaporation, supercooled

ground fog formed in a similar manner may be modified and even dispersed
if care is exercised to prevent overseeding.

In order to disperse a cold fog of this sort, it is necessary to use up the
available condensed water by seeding with only enough ice nuclei that the
crystals grow large enough to precipitate. An average concentration of about
20 ice nuclei per cubic centimeter is about the number required to produce

this effect.
If higher concentrations are used, there is a real danger that the density

of the fog will actually increase, thus reducing the visibility to a remarkable
degree. Since most ground fogs rarely contain more than 200 particles per
cubic centimeter, it is a simple matter to produce 10,000 per cubie centimeter
of ice crystals in the same volume of air. This not only reduces the visibility
but also makes the fog considerably more stable due to the very small particle

size and the further removal of moisture from the air.

Very peculiar optical effects occur in an overseeded cold fog, The outline
of objects near the limit of visibility become extremely fuzzy and of an unreal
appearance. In addition, the light scattered in the fog has a peculiar bluish
cast due to the Rayleigh scattering from particles small with respect to the
wavelength.

The prevention of the formation of ice fog is another possi
be attained by the proper manipulation of seeding technique:
an optimum number of sublimation nuclei into the a
fogs are troublesome, it may be possible to continuously remove the moisture

ility that may
By introducing
ions where such

Vol. 1, 1950 Experimental Meteorology 221

from the air which is responsible for the formation of this interesting but often
troublesome type of ground fog.

The ice crystals generated in the vortices of airplane propellers plus the
moisture added to the air by the combustion exhaust of the plane are the effects
which generally lead to the formation of ice fogs at airports.

Whether the removal of supersaturation with respect to ice by seeding
methods will be of sufficient magnitude to prevent ice fogging effects produced
by plane operations: can be determined most conclusively by actual experi-
mentation.

The modification of icing clouds for the protection of aircraft

This effect suggests an interesting possibility—the elimination of icing
clouds in the vicinities of airports and along heavily traveled air lanes. There
is no question about being able to accomplish the modification. The problem
which exists at present is whether or not it may have a practical application.

Low clouds which restrict the visibility for landing approaches around
airports, thick clouds in which planes must cruise as they wait for permission
to land, and thick clouds which might deposit a serious icing load on a plane
as it tries to climb up through them—these comprise hazards to safe plane
operations. Whenever such clouds are supercooled, they may be profoundly
modified as shown in Figs. 21 to 24. The practical and economic importance
of such operations can only be determined after detailed studies are made
in regions where such problems are thought to exist.

The simplest means for carrying out such cloud modification operations
is to employ a plane well equipped for flying under serious icing conditions.
Such a plane would be assigned the job of patrolling the air lanes, reporting
weather and cloud conditions and whenever serious supercooled clouds occurred,
would carry out seeding operations. A more direct means for protecting indi-
vidual planes may be the use of projectiles for modifying the cloud directly
ahead of a plane. This hardly seems practical for peace time operations, how-
ever, since a considerable hazard is involved in shooting anything into clouds.

Perhaps the most serious limitation to this use of cloud modification is
the fact that at the present extent of our knowledge icing clouds are nearly
unpredictable. The indefinite persistence of such clouds because of their
unstable nature is the feature that will probably prevent any effective use of
this application of cloud seeding within the near future.

In flying through a supercooled cloud, the airplane itself may produce a
fairly effective modification since the vortices which form at the trailing edge
of the wings and particularly from the propeller tips form large numbers of
ice crystals as the expanding air in the vortex cools below —39° C. Laboratory
studies of this effect indicate that as many as 10!” nuclei per cubic centimeter
222 Vincent J. SCHAEFER ZAMP

Vol ty tae Experimental Meteorology 223

may be formed in this manner. It is significant that when aircraft icing studies
are carried out in supercooled clouds, it is difficult to obtain an accumulation
of ice on the plane by making successive passes through a particular cloud.
‘After the first traverse, the icing property of the cloud is radically changed,
and it is often impossible to find any supercooled cloud in a region where
heavy icing was present a few minutes earlier.

With towering cumulus, however, having a vertical thickness of 5 km and
1) average liquid water content of 3 g/m, the precipitable water is more
ihn !y inch. Since such clouds are potentially unstable, the sudden conver-
jon (o rain might also lead to a local lifting of moist air which could produce
evel More precipitable water. Such clouds may often be seen to form and
(iosipate without producing any precipitation. It is for this reason that careful
{ilies should be conducted to learn everything possible about the physical
id colloidal properties of towering cumulus.

The modification of orographic thunderstorms

‘An extremely important type of orographic cloud is the towering cumulus.
s, certain peculiar topographic features combine

In most mountainous region The modification of towering cumulus
to favor the local formation of clouds. At certain seasons of the year, the
clouds generated by these “cloud breeding” regions develop into such large
cumulus that they become thunderstorms. Such storms once formed often
become detached from their site of development and produce violent distur~
bances. This aspect of these clouds will be discussed in a later section.

Since orographic cumulus formations are common in specific regions (21),
they provide nearly ideal conditions for research studies and the evaluation
of various techniques in experimental meteorology.

It may be possible that silver iodide seeding from ground generators woul
be particularly useful in modifying orographic cumulus to prevent their growth
into thunderstorms, By determining the air trajectory from the ground int
the cold part of the cloud, potential ice nuclei may be sent aloft by a ver
simple procedure. Since silver iodide particles become quite effective in th
region between —12°C and —16°C at which temperature the largest diffe

\s indicated in the last paragraph, the cloud structure of great economic
liiportunee is the towering cumulus. While such clouds often are produced
hy orographic lifting, they also form over flat country at times when the
iHosphere is conditionally unstable. Differences in ground heating and con-
(acl effects between warm and cold air along frontal systems often lead to

the formation of large regions of such cloud structures. If local conditions
permit the continued development of such storms within two to fiv
(hey may develop into thunderstorm:

hours
Dangerous and often deadly lightning
(rokos, torrential rains, destructive winds, and sometimes hail and tornadoes
Wye the end products of such development

Invariably such storms in their formative stages are characterized by a
Hil quid water content, strong vertical velocities, and supercooled clouds
Whose vertical thickness may exceed 5 km before many ice crystals form.

sien octets between the partial vapor pressures of water and ice, it is quit Vi large volume of supercooled cloud is invariably observed during the
possible that such clouds could be profoundly modified by permeating theif | ee {a thunderstorm and must, to a large extent, be responsible
general area with effective ice nuclei, If subsequent experiments indicate thal f len outbreak of such storms once their growth exceeds a critical

liye. The lange degree of instability due to supercooling may be altered
Within lew minutes as ice crystals invade the cloud. The large amount of
pyoigy released in this process provides additional impetus to the vertical
development of the storm, The rate at which this shift in phase takes place
js one of the important variables determining the subsequent progress and

it is important to seed such clouds at a temperature only a few degrees colde
than the freezing point (0° C), it may become necessary to use dry ice dispense
from planes or carried into the clouds by free balloons or projectiles.

The pioneer work, however, must be accomplished using, if possible, bot
ground observation stations and aircraft. Aircraft alone must be used if th
nature of the region precludes the use of ground stations.

development of the storm. Since under some conditions, the increase in tem-
perature alone may exceed 3°C, the total amount of energy released in this
janner i of tremendous magnitude. In a relatively small storm, it may be
fjuivalent to that released by several atomic bombs. :

It is the presence of thick supercooled clouds which raise
jumelbility that profound changes may be induced in cloud

The potential precipitation contained in clouds

From quantitative considerations, it is obvious that unless convergence
moist air occurs during the development of clouds and precipitation, th
amount of rain or snow that may reach the ground is at best of minor import
ance in relation to the water resources of the earth. For example, a clow
having a vertical thickness of 2 km and an average liquid water content
1 gim? would produce only about 2 mm of rain if completely precipitate:

the distinct
tems which

wre growing into thunderstorms.

Since the high, vertical thickness of a supercooled cloud seems to be a basic
/equiaite in the formation of a thunderstorm, it may be quite feasible by proper
voding, methods to prevent this phase from developing. :

Ved 1, Wun Experimental Meteorology 225
224 Vincent J. SCHAEFER a

‘The manner in which this seeding is done may produce a wide variati Vhe upparent limitations to the modification of cloud systems

in the end results obtained. By seeding each cumulus tower with large numbe Aa’: ity: of thieephgisicdl, Shenomens,; thecarare defittte limilations in ithe

of crystals shortly after it rises above the freezing level, the cloud would DMM gyn ji which experimental meteorology may be employed in modifying

continuously dissipated and no extensive regions of supercooled cloud coulff (jy |i the free atmosphere. Some of these apparent limitations may die

develop. Appear ax our knowledge increases although most of the restrictions now recog-
On the other hand, it might be desirable to’ seed such clouds to TealiZM iyo) yy imposed by known physical laws.

the maximum possible energy release. This presumably would involve seedi Horemost of these is the factor of cloud size and type. Certain clouds, such

each cumulus tower just previous to the point of its maximum development yy (jy (4\) weather cumulus (cumulus humilus), have such a small volume

If this could be done effectively, it might be possible to build the storm in

il wil outricted area that, even though they are easily modified when super-
a much larger one then would develop under natural conditions.

fooled, their total liquid water content is inconsequential. As pointed out in

| joyous section, even clouds of considerable vertical thickness contain but
‘lalively small amounts of condensed water. Another complicating factor is
ie prenaiva. af ball by veoding wb tha the alr below larger clouds is sometimes so dry that a considerable amount

Of preripitition evaporates before it reaches the ground.

Of considerable economic importance is the possibility that hail stor Another type of cloud which is difficult to modify is the warm ground fog
might be prevented by seeding techniques. Hail is believed to form undé@fiyjjjjq) jy sactiation or advection. Such fogs are often extensive and of con-
conditions of strong vertical convection in cumulus clouds having a higMfaidejyhle economic importance, especially from the standpoint of airplane
liquid water content but low concentrations of ice nuclei. With relativelf&(yy((\¢ ¢qy trol, The natural structure of a fog precludes any simple method
few active nuclei within supercooled clouds at temperatures between —5° Gj yyulifyiny it. Generally, the vertical thickness is not more than a hundred
and —20°C, hail particles may grow very rapidly by the difference in thiHyyajoyy or yo with a cloudless sky above. This rules out the modification from
partial vapor pressures of water and ice aided by the agglomeration of relaffyjjyjyo hy forming precipitation in higher clouds to “rain out” the fog. On

tively large supercooled water droplets. {he other hand, supercooled ground fog may be modified and, in some cases,

If large numbers of ice nuclei were present, the competition for availabl \\yjojned by the intelligent application of presently known methods of infect-
water would be so keen that no particles could grow very large. The importan@{ijiy ()ye y(joxphere with seeding agents.
of silver iodide for this type of modification is obvious since it might be pr: \iiother weather situation where no method of relief is now apparent is
tical to use ground generators positioned in such a way that these sublimatiolf&y) (jy (ye of drought. This condition generally results from the stability of a
nuclei would be carried into the clouds by entrainment. Since Miiples Weather pattern in a manner which, at present, is not very well

particles become quite active at temperatures colder than —10° C, this sul
stance should be quite valuable for this application if a reliable techniq
can be developed to get the nuclei into the critical regions for it to
effective.

(wlertood, Drought is generally accompanied with either cloudless skies or
Hy clouds of small vertical and horizontal development due to strong inver-
(onw or thick layers of dry air.

\ (ypical drought condition of this type occurred in New York state be-

A considerable amount of basic information is needed on the various pr woo) Jijne | and June 25, 1949, A dome of high pressure of great stability
perties of storms that produce hail. In some parts of the country where sevelf\eyeyjod1 over the northeastern part of the United States and resisted the
hail damage is frequent, storms are formed over certain mountain ridges amlMyeyye)iinyy movement of a cold front to the westward. This high became
peaks that serve as cloud breeders. Such clouds should be particularly suit famiant for nearly a week and then slowly moved eastward causing the per-
for modification by ground generators since the air trajectory is definiteljajq( yyovenent of fairly moist and very warm air from the south. Cumulus

related to the flow of air up the mountain and into the clouds. fide occurred with increasing frequency after a week of nearly cloudless

To accomplish the desired results, it may be necessary to build up a comff\joy jy\\) wore restricted in their vertical development by a layer of dry air
centration of nuclei of the order of not less than 100 per cubic centimet@ s\a(\jy between 8,000 feet to 16,000 feet where the temperature ranged from
in all the air likely to be involved in forming the cloud. With efficient gem@My yo ( |) 6° C and the mixing ratio fell from 6 g/kg to less than 1 g/kg of
rators, this should be possible using approximately 100 mg of silver iodidf{\\y_ J )jo cloud structure was mostly of a diurnal nature with the nights being

for one cubic kilometer of air.
226 Vincent J. SCHAEFER ZAMP

cloudless, while the greatest development occurred toward evening due to the
convection produced by the sun.

On the evening of the fourteenth day, a few widely scattered and very
local showers occurred coming from single cloud systems. Fig. 25 illustrates
a cloud which produced a brief shower as it passed across a strip about two
miles wide by ten miles long. The cloud dissipated in about an hour, the aver-
age amount of rainfall within this area being considerably less than 0.05 o
scarcely enough to lay the dust. If a number of clouds of this kind formed
rain in succession, the accumulated moisture might be of importance. Unfor-
tunately, the development and dissipation of this cloud required nearly two
hours and eventually, reduced the cloud cover to zero since the heating effect
of the sun needed for this type of cloud formation was no longer present due
to the late hour of the day.

The development of convergence is an important feature in the formation
of appreciable amounts of rainfall in many parts of the world. As a rule, such
developments are generally accompanied by the occurrence of natural pre-
cipitation which continues so long as the convergent movement is present.
About the only thing that artificial modification of clouds might do under
such atmospheric conditions is the initiation of the precipitation cycle a few
hours before it would start naturally or, under some conditions, to delay the
onset of precipitation by overseeding.

An interesting and valuable analyses of the relationship of cloud types
and systems and the possible effect which might be produced in them by arti-
ficial seeding operations has been presented recently by BERGERON (22).
Papers of this kind are of the utmost value, especially when they consider
and evaluate the results of laboratory and field operations. It is the feeling
of the writer, however, that such evaluations must be limited at the present
time to a consideration of the cloud systems in regions of the world familiar
to the observer. Generalization without observational data may raise obstacles
which are not truly valid.

A series of experiments [23] carried out in Ohio by a group associated with
the United States Weather Bureau have reported results which they have
interpreted as of doubtful economic importance. A study of the results which
they describe could be interpreted with a more optimistic viewpoint as con-
firming many of the claims made thus far by other workers in the field.

In view of the present relatively crude techniques and rapid advances
now being made in the field of experimental meteorology, it may be wisdom
to refrain from making world-wide generalizations until more experimental
and observational data becomes available.

The research related to experimental meteorology which is underway in
such places as Australia [24], South Africa [25], Hawaii [26], Canada (27), and
Honduras (28) are typical examples of the attitude which is necessary to gain

Vol. 1, 1950 Experimental Meteorology 227

a proper perspective of the possibilities and limitations of cloud modification
in various parts of the world.

Acknowledgments

The size and scope of many research projects at the present time makes
it difficult, if not impossible, to adequately define and give credit to those
responsible for the success of a project. A well coordinated team of enthu-
siastic workers becomes more and more a necessity. Because of their inter-
related responsibilities, it is nearly impossible to place credit where it is due.

This is particularly true with respect to Project Cirrus which enjoys the
cooperative effort of members of the Army, Navy, Air Forces, as well as the
General Electric Research Laboratory. :

REFERENCES
{1] BercEron, T., On the Physics of Clouds and Precipitatio ém. Union géod. gé
Benen 2 i ds and Precipitation, Mém. Union géod. géo-
2) Jones, A., and Lewis, W,
Equipment, Tech. Note.
(3] Konter, H.,

Recommended Values for the Design of Ice Prevention
1855, National Advisory Committee for Aeronautics (1948)
An experimental Investigation on Sea Water Nuclei, Roy. Soc. Sci.

Upsala 72, No. 6 (1941).
Laspaane, H., Atmospheric Condensation Nuclei, Ergebn. kosm. Physik 3, 155-252
Woopcoc:

A., Sampling Atmospheric Sea Salt Ni 0 s
Me Ree ee ee tos, ea Salt Nuclei over the Ocean, Sears Found.

4) ScHarrer, V. e De o e J. Meteor
Scwari i ol The Detection of Ice Nuclei in the Free Atmosphere, J. Meteorol. 6:
cut, B., The Nucleation of Ice Formation by Silver Iodide, J. App! s. 18,
c WY bh Phy
(1947); Nucleation of Supercooled Water Clouds iver Iodide es
Sov es Sub Hes y ds by Silver ide Smokes,
6) SmitH-JoHannsen, R., Some Experiments on the Fi
See is e Free.
Dorsey, E. Ernest, The Freezing of S ed Water, Trans. F Soc
4 ‘ zing of Supercooled Water, Trans. Amer. Phil. Soc. 38,

(7) Perrerssen, SveRRE, Weather ' . —_
seis , Weather Analysis and Forecasting (New York and London,

[8] Lancmure, I., The Production of Rain b
I 0 ya Chain Reactio a
Temperatures above Freezing, J. Meteorol. 5175-192 (1048), waits Clouds at
[9] ScHAEFER, V. J., Report on Cloud Studies in Puert
9] R, V. J., Repe foud Studies Rico, Occasional Report No. 12
Project Cirrus, General E. pseaveh, Laboratan Pew Were
Project Cirrus, ‘General Electric Research Laboratory, Schenectady, New York
[10] Veraarr, A. W., Meer Zonneschijn Ii yf Y
BRAART, Zonneschijn In Het Nevelig Noorden, Meer Regen In De Tvo
WN. V. Seyffardt’s Bock en Muzickhandel, Amsterdam, 1931) ° e 1" >? 779Pe™
(11] Garumax, Lovts, Rain Produced at Will (Chicago, Mlinois, 1891)
12] FInDEISEN, W., Colloidal Meteorological Proces
Precipitation, Meteorol, Z. 55, 121-133 (1938)

zing of Water, Science 108,

‘ses in the Formation of Atmospheric

228 Vincent J. SCHAEFER ZAMP

13] Apams, J. M., The Origin of Snowflakes, Phys. Rev. 35, 13-14 (1930).

14] Scnarrer, V. J., The Production of Ice Crystals in a Cloud of Supercooled Water
Droplets, Science 104, 457-459 (1946); The Formation of Ice Crystals in the Laboratory
and the Atmosphere, Chem. Rev. 44, No. 2, 291-320 (1949).

15] Scuarer, V. J., The Natural and Artificial Formation of Snow in the Atmosphere,
Trans. Amer. Geophys. Union 29, 492-498 (1948); Methods of Dissipating Clouds in
the Natural Atmosphere, Institute of Navigation (September~December 1947).

16] Scuarrer, V. J., The Production of Clouds Containing Supercooled Water Droplets
or Ice Crystals, Bull. Amer. Meteorol. Soc. 29, 175-182 (1948).

) Lonetey, R.W., On the Energy in a Hurricane, Bull. Amer. Meteorol. Soc. 30,

194-195 (1949)

18] BLANCHARD, D. C., Observations on the Behavior of Water Drops at Terminal Velocity
in Air, Occasional Report No. 7, Project Cirrus, General Electric Research Labora-
tory, Schenectady, New York (1948)

19] Conover, JonN, and Wotzaston, S. H., Cloud Systems of a Winter Cyclone, Blue
Hill Observatory Records. Presented January 28, 1949, at the New York Meeting
of the American Meteorological Society

20] Lanemurr, L., The Growth of Particles in Smokes and Clouds and the Production of
Snow from Supercooled Clouds, Proc. Amer. Phil. Soc. 92, No. 3, 167-185 (1948).

[21] Scuagrer, V. J., The Possibilities of Modifying Lightning Storms in the Northern
Rocky Mountain’ Region, Occasional Report No. 11, Project Cirrus, General Electric
Research Laboratory, Schenectady, New York (1948).

(22] BercERon, T., The Problem of Artificial Control of Rainjall on the Globe, II, Tellus 7,
No. 3, 32-43 (1949).

23) Coons, R. D., GENTRY, R. C., and Gunn, R., First Partial Report on the Artificial
Production of Precipitation, Stratiform Clouds, Bull, Amer. Meteorol. Soc. 29, 266
(1948); Second Partial Report on the Artificial Production of Precipitation, Cumuliform
Clouds, Bull. Amer. Meteorol. Soc. 29, 544 (1948).

[24] Kraus, E. B., and Sourres, P., Experiments on the Stimulation of Clouds to Produce
Rain, Nature 159, 489-494 (1947).

Sourres, P., and Smitn, E. J., The Artificial Release of Precipitation by Means of
Dry Ice, Unpublished report received from Australian Embassy, Washington, D. C.,
20 pages, 2 figures, and 2 tables.

Situ, E. J., Five Experiments in Seeding Cumuliform Cloud Layers with Dry Ice,
‘Austral. Sci. Res. [A], Phys. Sci. 2, No. 1, 78-91 (1949).

25] Anox, Artificial Stimulation of Precipitation, Interim Progress Report on Experi-
ments Carried out by the Council for Scientific and Industrial Research, the Division
of Meteorology, and the South African Air Force, Series Reeks M 551, 578, 9 (Pre-
toria, 1948)

(26) Leopon, L. B., and Hatsteap, M. H., First Trials of the Schaefer-Langmuir Dry
Ice Cloud Seeding Technique in Hawaii, Bull. Amer. Meteorol. Soc. 29, 525-534 (1948).

27) Fraser, D., Induced Precipitation Preliminary Experiments on Seeding a Variety
of Clouds with Dry Ice, Report MR-4, Div. of Mech. Eng., National Research Council
of Canada (Ottawa, January 1949),

28) Turnsutt, W., and SILvERTHORNE, J., Private communications. Results to be pub-
lished soon by I. LANGMUIR.

(Received: October 1, 49.) Printed in Switzerland,

Vol. 1, 1950 Experimental Meteorology

Fig. 4

Photomicrographs of stellar ervstals formed on cirrus type hexagons

229

230 Vincent J. SCHAEFER cAMP $8; 1, 1080.

231

Fig, 15, Portion of 15 miles long “‘L” pattern produced with about 1-1 pounds per mile of dry
ice which resulted in overseeding effect. Note reflection. Flight 3, April 7, 1947, Project Cirrus.
Official Photo, Signal Corps Engineering Laboratory.)

Fig. 16. Remnants of six separate spot drops of 1-¥%4 pounds of dry ice. These six ice crystal clouds Fiz. 17. A-view of the “Gamma”

n : Famma'” pattern after 21 minutes produced by seeding with dry ice pellets
persisted after all other supercooled clouds dissipated. Flight 23, April 29, 1948, Project Cirrus. it the rate of 1-3 pound per mile. Each leg (not all shown) Lea nie on silts Hy Ne re
Official Photo, Signal Corps Engineering Laboratory.) 24, 1848, Project Cirrus. (Official Photo, Signal Corps Engineering Laboratery.) en ne
232 Vincent J. SCHAEFER zame Vol. 1, 1950 Experimental Me

Fig. 19. View of ‘Figure Four" pattern produced by dry ice seeding. Legs 10 miles long. Flight 73
March 10, 1949, Project Cirrus. (Official Photo, Signal Corps Engineering Laboratory.)

“ ” patter 3 y ice seeding using less than one pound of
Fig. Is, View of “Racetrack” pattern produced with dry ice seeding using u . .
dy ice per mile of flight. Straight legs are 18 miles long. Flight 53, November 24, 1948. Project Fig. 20, View of two ‘L” type patterns produced by dry ice
Cirrus. (Official Photo, Signal Corps Engineering Laboratory.) Flight 80, March 31, 1949, Project Cirrus, (Official Photo,

eding of supercooled stratus cloud

al Corps Engineering Laboratory.)

234 Vincent J, SCHAEFER zAMP

Experimental Meteorology 235

Fig. 23. Hole having an area of 70 square miles cut into a stratus cloud by a dry
“LY pattern. Flight 23, April 29, 1948, Project Cirrus. (Official Photo, $
Fig. 21, View of a large gap in an overcast produced by dry ice seeding with new clouds beginning | aboratory.) rh
to fill in the empty spaces. Flight 52, November 24, 1948, Project Cirrus. (Official Photo, Signal

Corps Engineering Laboratory.)

seeding in an
ignal Corps Engineering

Fig. 22, Portion of nearly 250 square miles of stratus cloud removed by s

Edlogattsis of founds hig. 24. Large area of supercooled stratus cloud modified with dry lee. Mliht Ni, Project Cirrus
per mile. Flight 52, November 24, 1948, Project Cirrus. (Official Photo, Signal Corps Engineering Official Photo, Signal Corps Engineering Laboratory.)
Laboratory.)

236 Vincent J. SCHAEFER zaMe

Fig. 25, Appearance of cumulo-nimbus cloud system which produced the first local showers on the

fourteenth day of a drought in Eastern New York

GENERAL@ ELECTRIC

\ i : : . e
we i
/ JN G Geferal News Bureau
| ]- oom 104
/ BLDG, #23 ‘

SUBJECT

a ld g reo

—<\/ : Schenectady, February 23, 1950
(Research Laboratory)

/
Mr. /G. W. Griffin, Jr.

Dear George:

Miss Alice Neil, Research librarian, celled my attention
to the accompanying proofs of an article by Vincent Schaefer
which embodies the material that he gave last summer before
the United Nations conference on the "Conservation and Utiliza-
tion of Resources, at Lake Success, It is to be published in
the ZUtschieft flr Angewandte Mathematik und Physik, Zirich as
br. KR. Singer, Editor,

Though the footnote states that this has been approved
by the National Military Establishment, neither Miss Neil nor
I have a record of any company approval. In view of the question
about this whole project, I am sending it to you for your
examination,

Dr. Schaefer has returned another proof to the editor
with his corrections, so it would be very difficult to make
any chenges at this time, however, I believe that this is more
restrained than the other article which we were discussing
and so there should not be any difficulties. I believe that
Vince is anxious to get this copy back so please return it to
him either directly or through me, as soon as possible.

Very truly yours,

5

James Stokley/jmw

re) 2 Oe chars oh Ose
eee

APPROVAL OF PAPER FOR PUBLICATION

Subject EXPERIMENTAL METEOROLOGY
Author Dr. V. J. Schaefer
Use Zeitschrifft ftir Angewandte Mathematik und Physik, Zurich

The attached manuscript has been submitted for approval of the
Company, as a paper for publication. Please make any changes or
corrections necessary and return promptly. If you make any changes,
please indicate the pages on this sheet.

Submitted Approved-Date

Changes on Pages

.. the record, It, has had Military, approyal..................
Comments:
RETURN TO

James Stokley
fe. Advertising and Publicity Dept.
Room , Bldg. #23
Phone 4610 Schenectady, New York

APPROVAL OF PAPER FOR PUBLICATION

Subject E XPERIMENTAL METEOROLOGY

Author Dr. V. J. Schaefer

Use Zeitechrieft fur Angewandte Mathematik und Physik, Zurich

ae paar bi ), Coke

The attached manuscript has been submitted for approval of the
Company, as a paper for publication. Please make any changes or
corrections necessary and return promptly. If you make any changes,
please indicate the pages on this sheet.

Submitted Approved-Date Changes on Pages

Comments:

RETURN TO

James Stokley
Yo Advertising and Publicity Dept.
Room 367, Bldg. #5 23
Phone 4610 Schenectady, New York