GENERAL @ ELECTRIC
Keewech Laboratory
HISTORY OF PROJECT CIRRUS
Compiled by Barrington S. Havens
Public Relations Services Division
Report No. RL-756 July 1952
CLASS 1
SCHENECTADY, NEW YORK
CLASS 1:
CLASS 2:
CLASS 3:
CLASS 4:
CLASSES OF GENERAL ELECTRIC
TECHNICAL REPORTS
Available to anyone upon request.
| Listed and abstracted
| in T. I. S. Briefs.
Available to G-E employees for use within
the Company.
Limited distribution within the Company. __
Rigidly limited distribution within the Company.
May not be sent outside the United States.
(Class 4 reports should be returned to Research
Laboratory Publications if they are not needed.)
GENERAL @@ ELECTRIC
Kecewuk Libovalty
REPORT NO, RL-756
HISTORY OF PROJECT CIRRUS
Compiled by Barrington 8. Havens
Public Relations Services Division
July 1952
Published by
Research Publication Services
The Knolls
Schenectady, New York
FOR USE OF G-E EMPLOYEES ONLY
GENERAL @@ ELECTRIC
Reseach Labowliy
SCHENECTADY, NEW YORK
TECHNICAL INFORMATION SERIES
Title Page
AUTHOR SUBJECT CLASSIFICATION No.
Havens, RL-756
q DATE
Barrington S. meteorology July 1952
TITLE Tapes ae
History of Project Cirrus
ABSTRACT
Project Cirrus, initiated on February 28, 1947 under
Contract W-36-039-sc-32427, requisition EDG 21190, was
established to cover ‘‘research study of cloud particles an
cloud modifications.’’ Project Cirrus continued through th
life of several government contracts, ending in 1952. A
history of the project covers not only the work done under
G.E. CLASS REPRODUCIBLE COPY FILED AT NO. PAGES
ae Seen Lae Research Publication
; Services 105
BSNCLUSIGNS° Sq o> sag Tai. Sk ee ee
history of the project covers not only the work done under
government contract but also the work of General Electric
scientists for many years leading up to the establishment
of the project.
By cutting out this rectangle and folding on the center line, the above information can be fitted
into a standard card file.
INFORMATION PREPARED FOR Research Laboratory
TESTS MADE BY.
COUNTERSIGNED DIV.
DIVISIONS LOCATION.
FN-610-IM-RL (2-50)
TABLE OF CONTENTS
Page
Tienes TIMES O GNC ELON wins leis ccs ono oo ae, ao, oi eons Rey eee eee ais iL
Ti = Warly HiStOry oocccencnc soc ecscecsenccsnescte see 3
Gas Masks & Smoke FiltersS ......0cesceerssoee
Smoke Generators ...-cecccerrcees Sistas altancuatiare)e
PYeCIPItTAtiON SUATIC 2. o> ees. POE A ceed ee Est Mae be
INU CSeNNe Whchbalee oie eh eror cco Gaeta coke lacelecetcyeienee oi a=te
Cloud Studies at Mt. Washington......-....6- shiv
INMEGICENICY GRA Ga pea connote anc cs ial atevatter ene ei
Schaefer’s Cold Box......-- Bese ai ray Tassel ee vat steel sie
Vonnegut’s Early Work - Cloud Studies au Moye an.
Shicemeoolhiie han csAmecook: Sos AIO ERC ee ROR os eats
Supercooling of Metals ....2c-cecscorserererees :
Nucleation StudieS ....cccesccecsercossrceeces :
Sbvers HOHE s 2 A Rowe gue oe ee, Ret eas ott Pelieyenltette
Langmuir’s Early Seeding Calculations ......---
First Man-made Snowstorm. ...+.--++«- coalesacuauenate
Other Early Flights....... Bee areata cae Re case etree aes
Establishment of Project CirruS ....-cescsesers
AMOI wW Ww
NF Oo oO @on7
ht
boo
o1
Ill - Getting Organized...... Sree Eade e te ee ve cayrerationeneuan ctererolabecars
—
o1
Contractual HiStory ..csccccccsccvscssceorce as
Organization ..... tA N eM arsitai ante easel Sihc\radeaifexereaeiee) meats shai
Flight Program. ..cccscssscccscessesseeererecs
G@roniad @pera HONS. 2 2s elas 6.2 cc oxe crocs cen e a0 cnie
WE SEWGIS Start OW is a edie en ets cue ucke slates eumiesaieia, = elelele)6
Photography.....- eRe haTiMae ay tol ct nesns ae nig oleveus a aneye cnt 2
IbavSwebasalsralgehulodl = 4 Aaa oc DE aun N Ie ace har i clionohsua vas
MNMRPRREH
FPOOMONMN
wo)
—
Dry Ice DiSpensSer. ...sccceccccccereoecrecce
Dry lee Crusher, wc > ccc 25-2 See ree MON Or
Silver Iodide Generators ...cccscresceecscere
@andera. Glim@mleter yiermte oc 2 ele als sons, eco éleleialelee
Flight Instruments ...cccccccrscseasocrscree
‘e\WWeather’’? InstrumentS .....cc2ccceccrrcsoes
GillouGdaieter:: cs Cobo nu Ue ODe ono
IX - Co-operation with Other ProjectS ........-.2+-+sse+-ee.
Pineapple Research Institute, Honolulu, Hawaii......
Milliken & Farwell, Mobile, Alabama...............
Uinitbacl emis Cormajoeiony; Islomelmrs 225 62556h556o5e505-
‘New Work City Water Shortage. 250. .2 2s eset eee
Commercial Seeding in the WeSt............--22-----
Worle oi Ouner Gowvemayaneias socsqccbosncsGoscoudomuo
K co ComeluSiOn soacocnovoscccon sooo Dood Gun eOnDOD oO GOnKON
OST URS SUES anos rice ueewaterecel cis teakaneione: ons keeey lage romsteyal«
Widespread Weather Modification ........-.-..0+6-
Modifying Orographic Clouds .......-..--+ss0+-e-
Producing Regions of Ice Nuclei................--
Modifying Stratiform CloudS..........-.++.sse+ee-
Modifying Supercooled Ground Fogs ........-+-+:.
PrQOUACSuloia, Cie JAVMCIGENES 6 also OU oOODo Kd OGD BOO UTCOou4
Modifying Orographic Thunderstorms .........-..
Modifying Towering Cumulus ..........+--+-e+eee:
Preventing Hail... 52. oct et ee ee eee sens
ANoVORNeSING ILgbadwAloris 5-6 Ga boacdcuOo noe oC Ou an Goo OG
Isehie Weenie (Clasbulblse 5 ag an sbo oo poe poonoedooDe
Wiereian GreOwmnIG ITOLS, saanchsoocsesoccdcdceoeednom
IDISOWIENI me eGo Ko Guu oc ouo coo Uden Ee Omeoeo Oooo one
(COIWIEOEMNOCO Scan scGoncdoécancco see OUuCOUD odor
Controversial ASpeetsieyemus ete eels « ateter et «aula aie ate tsifelle)(ail)(oll=
ILSCISENMIOM A Boa nob oaos Goueo loo doGomos oUC MOR ooo oE
SHS EMC CwLORALUITS 55-2 ela) synicue sic eiaiolercus ela)» loi cile le) win lice) 6
Appendices
I
II
III
IV
Alphabetical List of Personnel ...........-.--
Hepsi Cala MIR SUS oye tesen eee atcltatel(ei'e letatiel = far -ueuey “cot -Jaiol shoxevore) =
Ground Operations..... SPR OT CLG RC Gc ONCE
Bibliography of Reference Literature .........
PROJECT CIRRUS HISTORY
I - INTRODUCTION
This history of Project Cirrus was prepared at the request of the
Research Laboratory for three reasons. First of all, the project has been--
and still is, at this writing--of such unusual interest and significance, that
the telling of the story is merited for its own sake. Secondly, the termina-
tion of the project is bound to result in an eventual dispersal of the various
members of its personnel. Already Dr. Langmuir has retired from active
General Electric employ, and the other members of the project are, and
will be, more and more engaged in new and completely different activities.
And finally, the broad aspects of the project have such wide implications that
it is particularly important that the story be committed to paper ‘for the
record’’,
It has not been easy to organize the raw material in any simple, logical
fashion. As is so often the case, the project was very complex, with a num-
ber of subdivisions associated with the main activity. Some of these subdi-
visions ran consecutively, some operated in parallel, and others intertwined
or branched off in variously divergent directions.
Where it was possible the material has been arranged in chronological
or otherwise logical order. Where it was not possible, the various subor-
dinate topics have been taken up in as nearly a logical order as possible. As
a result, cases will be found where the story “‘gets ahead of itself”’, and
later it becomes necessary to retrace one’s steps to pick up the thread.
The history, with the exception of the Introduction and Conclusion,
divides itself naturally into two main parts. The first is the story of the
early activities which led to the formation of Project Cirrus. The second
is the story of Project Cirrus itself.
Schenectady, New York — B.S. Havens
July, 1952
Il - EARLY HISTORY
It would be difficult, if not impossible, to trace the complete lineage
of everything leading up to Project Cirrus. General Electric scientists were
not the only ones who studied many of the problems involved. And even when
restricting consideration to General Electric research projects, the situa-
tion is complicated.
The following material is confined as much as possible to work which
has a relatively direct bearing on Project Cirrus research.
GAS MASKS & SMOKE FILTERS
The earliest activity leading directly to Project Cirrus was the study,
beginning in 1940, of the fundamental nature of filtration in gas masks. This
work was undertaken by Dr. Irving Langmuir ens Dr. Vincent J. Schaefer at
the request of the Chemical Warfare Service. 12)
‘Gas masks normally use charcoal to absorb poison gases, but even in
World War I the possibility arose that the enemy might use toxic smokes
which could not be absorbed by charcoal and thus would have to be removed
by.a filter somewhat like filter paper.
The first step in attacking the problem was to make some smokes of
the type for which the filters would be used. In doing so, the scientists stud-
ied the particles which composed the smokes. They investigated such things
as particle stability, concentration, and measurement. They obtained fairly
successful theoretical results and a better understanding of how to build a
good filter. And incidentally, they acquired a great deat of detailed know-
ledge as to how to make a smoke which would be non-volatile and would con-
sist of particles far smaller than those of ordinary smokes, and they learned
much about optical properties.
This work was done under a National Defense Research Committee
contract. As Langmuir and Schaefer neared the end of the work, a form
letter was received in August, 1941, asking if anyone could think of a way
to make a white screening smoke that could be used over large areas to
cut down the hazard from aerial bombardment.
SMOKE GENERATORS
Langmuir and Schaefer wondered whether they couldn’t do this by
using the methods they had adopted for making smokes for testing filters.
They decided to try.
They had found that the easiest way to make smokes and control the
particle size was to take some oil and put it into a volatile condition. They
Early History -4-
heated oleic acid and similar substances up to about 200°C and passed a
stream of air over them to get the vapor mixed with air. Then they quenched
the mixture suddenly by blowing in a large amount of cold’ air. The parti->
cles grew in size and by sudden quenching they found they could stop the
growth at any desired point and also make particles of very small size.
They were surprised to find that, under certain conditions, they could get
particles of extraordinarily uniform size.
Further work and experimentation showed thatthey could do the same
thing on a large scale. Larger generators were built, tests were made, and
the design was adopted by the Army and used successfully and on a large
scale curing the war. :
PRECIPITATION STATIC
Quite independently of this work the Secretary of War asked in 1948
for research into the problems of precipitation static.(12) It was believed
that the invasion by Japan would have to come very largely from air attacks
through the Aleutian Islands, across Alaska, and from the North. That led to
a tremendous development of air transport and airplanes through the Aleutians.
The difficulty in flying aircraft in the Aleutians was very serious. One
of the big problems was icing of the aircraft, but even more baffling was the
complete loss of radio contact when the planes flew through snowstorms. The
planes might become charged, sometimes, to a potential of 250,000 volts or
more, producing corona discharges from all parts of the plane and causing such
electrical disturbances that radio sets could not receive messages. Pilots had
particular difficulty in finding their bases and getting down through this foggy
bad weather. What could be done about it?
Langmuir and Schaefer were interested. They had no particular ideas
on the subject, except that it had to do with weather. In their opinion, the best
place to investigate something like that was the well-equipped laboratory of the
Mt. Washington Observatory on top of Mount Washington in New Hampshire.
Mount Washington in winter has an average temperature of minus four
or five degrees F, the wind averages about 60 miles per hour, and most of the
time clouds sweep over the summit. It seemed to offer the proper conditions -
for a research of this kind.
So equipment was installed at the summit, and Schaefer went there. sev-
eral times during the winter of 1943 to conduct experiments. But he discov-
ered that anything exposed there during the winter immediately became cov-
ered with ice, because the air was full of supercooled water droplets. He and
Langmuir became So much interested in this that they hoped they would not
have to continue a long study of precipitation static.
Early History aie
In the course of this work, Schaefer relied heavily on the services
of Raymond E. Falconer, who was then one of the observers in the weather
station on the summit.
AIRCRAFT ICING
It So happened that the Army Air Forces were just as much interested
in problems of aircraft icing as in precipitation static. This fitted in so well
with the new interest of ae and Schaefer that in 1944 they starteda
study of icing of aircraft. 32A)
They had much assistance from Victor Clark, Falconer, and others
of the observatory personnel, who were already working on riming and
icing. Langmuir and Schaefer, however, were able to introduce Some new
and very productive ideas.
Extensive mathematical calculations were necessary. The first work
of this nature was done by Langmuir, and his results were used in connec-
tion with the cloud studies at Mount Washington (see below). During the
later stages of the Mount Washington studies, Langmuir decided to make use
of a differential analyzer for these calculations, and in preparing the mat-
erial for that purpose, he was assisted by Dr. Katharine Blodgett. Thus
it was possible to calculate the percentage of water droplets which would
be depesited on a given surface under specific conditions, The information
was used on data obtained on Mount Washington to determine the number
and size of water droplets involved in the formation of ice.
CLOUD STUDIES AT MOUNT WASHINGTON
The theoretical calculations worked beautifully in practice. They
began to acquire a very satisfactory understanding of some features of
cloud structure and the growth of cloud particles. They became absorbed
in this new interest. And Langmuir found he could apply to his smoke gen-
erator work the same evaporation-condensation theory he had used to cal-
culate the growth of smoke particles.
But, although they felt they had a fundamental theory for some of the
factors that caused particles to grow in clouds to the size they are, they
didn’t feel conditions were right for further study on Mount Washington.
It would be far better to study cloud particle growth in airplane flights.
That would require the development of new instruments.
This was late in 1946. They took the question up with the Army Air
Force and the Signal Corps. They were led to think that perhaps some-
body might furnish aircraft for experimental purposes of this sort; it
seemed that it would be desirable to know something about clouds from a
Early History -6-
standpoint of national defense. But they didn’t get along very fast. They
carried the research along on their own to a large extent, testing instruments
on Mount Washington, but they never got tests in aircraft.
NUCLEATION
By this time they were deeply interested in their cloud study. They
investigated and learned a lot of things. But the thing that struck them most
was that, if there are any snow crystals in a supercooled cloud, they must
grow rapidly and should tend to fall out. They came to the conclusion that
in winter, if there are supercooled stratus clouds from which no snow is
falling, even though the temperatures in the clouds are below freezing, there
simply are no appreciable numbers of effective snow nuclei. Such clouds can
apparently be supercooled to very low temperatures.
They thought this presented a problem that should be investigated. Why
was it that sometimes snow forms so easily, with apparently no lack of nuclei
on which crystals can grow, and at other times there seem to be none? They
concluded there must be something in the atmosphere that causes water drop-
lets to change to ice only at certain times and under various conditions. They
decided to make some careful experiments in the laboratory in an attempt to
duplicate those conditions.
SCHAEFER’S COLD BOX
During Langmuir’s absence in California for three or four months in 1946, |
Schaefer made what Langmuir has described as ‘‘some beautiful experiments? (1!
During the previous winter he had been studying the behavior of droplets on cold
surfaces to see how they supercooled or froze as the temperature dropped. He
had found he could supercool water drops to as low as -20°C on surfaces coated
with polystyrene and similar materials. He had realized, however, that such
experiments were not simulating supercooled clouds and had sought a better
method of experiment.
He decided to try a home freezing unit of the type used for food storage.
He lined it with black velvet so he could get a good view of what happened inside
when he directed a beam of light down into the box. He then breathed into the box,
and the moisture condensed and formed fog particles which were just like ordin-
ary cloud particles, although the temperature was about -23°C. No ice crystals
formed. He tried many different substances dusted into the box to get ice crys-
tals to form, but almost never got any. He got just enough to convince him that,
if he did get them he could easily see them.
Finally, one July day when the temperature of the chamber was not low
enough, he put a big piece of dry ice into it to lower thetemperature. In an
instant the air was full of ice crystals. The crystals persisted for a while
Early History ae
after he took the dry ice out.
Following this discovery, Schaefer conducted a number of experiments.
These showed that even a tiny grain of dry ice would transform the super-
cooled cloud in the cold box to ice crystals. Quantitative experiments were
conducted which showed that many millions of crystals could be produced
in this manner.
In order to find out if there was something peculiar to dry ice which
produced this effect, he worked with other cold materials. For example,
he showed that, by dipping a common sewing needle into liquid air and then
passing it momentarily through the supercooled cloud in the cold box, sim-
ilar spectacular effects occurred. This demonstrated that the presence of
a sufficiently cold substance was all that was required to produce the effect.
Schaefer devised methods and equipment for determining, with considerable
accuracy, the fone temperature at which the supercooled cloud changed
to ice crystals. 36) This temperature was found to be -38.9C+0.1 degree.
Schaefer’s discovery changed the whole situation. It meant, first, that
it was not the dry ice or the needle as such that was responsible for the ef-
fect, but the temperature. Anything could be used having a temperature of
-40-C or colder.
VONNEGUT’S EARLY WORK
CROUDSTUDIES AT Vela.
Meanwhile the stage had been set for another important contribution
to this pioneering work in meteorology. Before Dr. Bernard Vonnegut be-
came associated with the General Electric Research Laboratory, he was
employed at Massachusetts Institute of Technology, where he had been en-
gaged in various stwiies during the early years of World War II, In the
laboratory of the Chemical Engineering Department he worked on smokes
for the Government’s Chemical Warfare Service. He measured smokes,
smoke penetration, and smoke filters. Then he became interested in the
problem of icing of airplanes and went to work on that in the Meteorology
Department, for the Air Force.
SUPERCOOLING
Meanwhile he had been doing some work on the side in supercooling.
He found that by making an emulsion of water drops suspended in oil, he
could cool water far below the normal freezing point, and it would not
freeze i a certain point was reached, when the whole mass froze very
rapidly. 62
Early History -8-
Vonnegut joined the staff of the Research Laboratory in the Fall of
1945 and he continued his supercooling investigations there.
SUPERCOOLING OF METALS
In various contacts with Langmuir and Schaefer, Vonnegut learned of
the work they were doing. Knowing that Schaefer was already working on the
supercooling of water, he switched his activity to the supercooling of metals,
in order to avoid duplication. He found he could supercool Woods metal by
subdividing it into many small, independent particles, and he developed a
technique of studying the effect with x-rays. He also-worked with tin, (62)
NUCLEATION STUDIES
Vonnegut had been interested in the work being done by Langmuir and
Schaefer and had kept in rather close touch with it. In the fall of 1946, Lang-
muir asked him if he would be interested in helping with the quantitative work
being done on the number of ice crystals produced by dry ice. As a result,
Vonnegut applied himself to this and other problems in the general study of
nucleation.
SILVER IODIDE
It occurred to Vonnegut that some substance very similar to ice in its
crystal structure might serve as the nucleus for the formation of ice crystals
in the cold box. He went through all the known tables of crystal structure and,
from over a thousand compounds, selected three substances that he thought
might have possibilities: lead iodide, antimony and silver iodide. 56
He dropped samples of each of these three substances into Schaefer’s
cold box. The results were almost negligible, although he produced enough
effect with the lead iodide to warrant further experiment. He and Schaefer
tried iodoform and iodine and obtained ice crystals in small numbers with
them, too, but nowhere near as many as with dry ice seeding.
The problem intrigued Vonnegut. He decided to try a metal smoke in-
stead of the powder. He introduced some silver smoke into the box by draw-
ing an electric spark from a piece of silver, and it produced in the cold box
a swarm of ice crystals.
The results were So spectacular that he decided to try silver iodide
again, but this time as a smoke, for the effect with silver did not persist.
First he vaporized silver iodide and then he introduced into the cold box
the smoke resulting from the rapid condensation of this vapor. It was a com-
plete success. Further investigation showed that his earlier negative results
Early History 29
with silver iodide had been caused by the fact that the silver iodide he had
used was impure. Powdered silver iodide worked very well when it was
reasonably pure. He also found that the reason for the successful use of
iodine was again impurity--contamination with silver.
The problem then became one of finding out something about how
silver iodide worked and of finding methods of generating silver -iodide
smoke of small particle size on a large scale. So many nuclei could be
produced with silver-iodide smoke that calculations indicated all the air
of the United States could be nucleated at one time with a few pounds of
silver iodide, so that the air would contain one particle of silver iodide
per cubic inch--far more than the number of ice nuclei occurring nor-
mally under natural conditions. 65
LANGMUIR’S EARLY SEEDING CALCULATIONS
Meanwhile Schaefer and Langmuir had continued their study of
the effects of dry ice. In August of 1946 Langmuir made a theoretical
study of the rate of growth of the nuclei produced by dropping pellets
of dry ice through clouds of supercooled water.(80) He calculated the
velocity of fall and time of dissipation of the dry ice, the amount of ice
particles that would be formed, their size, the amount of snow which
would result, etc. With a reasonable number of pellets dropped along
a flight path into the top of a cloud, the limiting factor would not be the
number of nuclei but the rate at which they could be distributed through=
out the cloud.
He also showed that such a formation of ice and snow particles
would raise the temperature of the cloud, and he calculated the amount
of temperature change. Thus the air in the cloud would be caused to
rise, increasing its upward velocity because of the seeding. The result-
ing turbulence would spread the ice nuclei throughout the cloud. He
anticipated that it would only be necessary to seed a stratus cloud along
lines one or two miles apart in order to give complete nucleation of the
cloud within'a period of 30 minutes or so.
FIRST MAN-MADE SNOWSTORM
Thus the stage was set for actual experiment with an airplane in
real clouds. On November 13, 1946, a Fairchild airplane was rented at
the Schenectady airport, piloted by Curtis Talbot, and Schaefer went
aloft in search of a suitable cloud. 38) It was found over Pittsfield,
about 30 miles east of Schenectady, at an altitude of 14,000 feet anda
temperature of -200C. What happened next is best described by the
following extract from Schaefer’s laboratory notebook entry for that
day:
Early History -10-
‘“‘Curt flew into the cloud and I started the dispenser in
operation. I dropped about three pounds (of dry ice) and then
Swung around and headed south.
‘‘About this time I looked toward the rear and was thrilled
to see long streamers of snow falling from the base of the cloud
through which we had just passed. I shouted to Curt to swing
around, and as we did so we passed through a mass of glistening
snow crystals!....We made another run through a dense portion
of the unseeded cloud, during which time I dispensed about three
more pounds of crushed dry ice..... This was done by opening
the window and letting the suction of the passing air remove it.
We then swung west of the cloud and observed draperies of snow
which seemed to hang for 2-3000 feet below us and noted the
cloud drying up rapidly, very similar to what we observe in the
cold box in the laboratory..... While still in the cloud as we saw
the glinting crystals all over, I turned to Curt and we shook hands
as I said ‘We did it!’ Needless to say, we were quite excited.
‘“‘The rapidity with which the CO, dispensed from the window
seemed to affect the cloud was amazing. It seemed as though it
almost exploded, the effect was so widespread and rapid........
‘“When we arrived at the port, Dr. Langmuir rushed out, enthu-
siastically exclaiming over the remarkable view they had of it in
the control tower of the G.E. Lab. He said that in less than two
minutes after we radiced that we were starting our run, long
draperies appeared from the cloud vicinity.”’
This first seeding flight was of tremendous significance. Not only did
it show that the laboratory experiments and calculations were justified, but
it also contributed new material to the rapidly accumulating store of know-
ledge. For example, it suggested that the veil of snow that first appeared
immediately below the cloud could not have been produced by snow falling
from the cloud but rather was produced directly by the action of the dry
ice pellets falling inte a layer of air below the cloud which was saturated
with respect to ice but not with respect to water.
Subsequent experiments proved that it was also frequently possible to
seed a supercooled cloud by flying just below it and dropping dry ice. The
thickness of the layer in which such seeding is possible is about 10 meters
for each degree C below the freezing point at the cloud base. The ice crys-
tals thus formed may be carried up into the cloud if the cloud is actively
growing by convection.
Early History -11-
On November 21 Schaefer seeded a supercooled valley fog with dry
ice. He found that it was possible to reduce visibility by generating more
ice crystals than fog droplets and also to dissipate the fog by dispensing
. just enough ice crystals to use up the fog droplets, each crystal growing
‘large enough to fall to the ground.
OTHER EARLY FLIGHTS
There were two other seeding flights made by Schaefer with a rented
plane that month, one on the 23d and the cther on the 29th. 79) These tests
were made on isolated cumulus-type clouds. The whole of each cloud was
changed into ice within five minutes, and snow began falling from the base
of the cloud.
Photographs were taken from the ground every 10 seconds, and these
were developed and projected as movies. They showed that, with orographic
clouds the air moves into one part and leaves another part; ina matter of
five minutes or So an entirely new mass of air is within the cloud. Thus
it was found that experiments with small cumulus clouds are usually of
little interest, for the effects last but a few minutes.
cigs flight test was made on December 20, also using a rented
plane. 2) This time the sky was completely overcast, and by 9 o’clock
in the morning the Weather Bureau in Albany reported that it expected
Snow by 7.0’clock that evening. Schaefer dropped about 25 pounds of
granulated dry ice in the lower part of the cloud at a rate of 1 to 2 pounds
per mile, about 1000 feet above the irregular and ragged base of the over-
cast, at altitudes ranging from 7000 to 8500 feet, at about noontime. A
two-pound bottle of liquid carbon dioxide was also discharged into the
cloud during this period.
Before and during the seeding flight, a light drizzle of supercooled
rain had been encountered, which seemed to evaporate before it reached
the ground. Flying back along the line of seeding, after seeding was com-
pleted, it was found that the drizzling rain had stopped and that it was
snowing. But on reaching the point where the seeding had stopped, drizzle
conditions were again encountered. Three more seeding runs were made
along the same line before the plane returned to Schenectady.
The plane then descended to 4000 feet, where the visibility was
better, and made a reconnoitering flight, checking the places where snow
was falling. By this method and through reports received, it was found
that snow started to fall in many places in the region. At 2:15 p.m, it
started snowing in Schenectady and at many other places within 100 miles.
It snowed at the rate of about one inch per hour for eight hours, bringing
the heaviest snowfall of the winter. While the seeding group did not
Early History -12-
assume it had caused this snowstorm, it did believe that, with weather con-
ditions as they were, they could have started a general snowstorm two to
four hours before it actually occurred, if they had been able to seed above
the clouds during the early morning.
ESTABLISHMENT OF PROJECT CIRRUS
This, then, was the situation in which the research workers found
themselves by the end of the year: Their work on precipitation static,
then on aircraft icing, had developed through cloud studies into meteoro-
logical work of profound significance. But, while their work on precipita -
tion static and aircraft icing had been done under government contract, the
work they were now doing on weather research was not. Their last con-
tract had expired at the end of the previous June.
At this point Dr. C. G. Suits, Director of the Research Laboratory,
reported some of the results of cloud seeding to General Electric officials.
While it was clear that weather modification and experimental meteorology
were remote from the research which had been the traditional interest of
the laboratory and the Company, it was equally clear that these new results
were possibly of very great significance to the country. It was, therefore,
decided that the work should be encouraged and pushed forward.
Because the results might have such wide application to the country
generally, and because much government assistance would be needed in the
form of weather data, airplanes, and flight equipment, a government con-
tract for the continuation of the work was to be sought. While the govern-
ment agency which had sponsored the previous research was not interested
in the new work, other government agencies were. Normal contacts with the
Signal Corps, for example, had kept that organization in touch with the new
research, and Col. Yates, chief of the Air Weather Service, had asked the
Company to submit a bid covering this work in the latter part of September.
A formal proposal covering cloud modification and cloud particle studies
was submitted to the Evans Signal Laboratory at Belmar, New Jersey
(a Signal Corps unit) on September 20. Meanwhile the weather studies were
being conducted at General Electric expense, although General Electric anti-
cipated no benefit resulting to the Company from the work from a meteoro-
logical standpoint.
The flight test of December 20 added a powerful stimulus to the Com-
pany’s negotiations with the government. Although the General Electric press
release covering it did not claim that the general snowstorm was caused by
the seeding, the coincidence of the two events did cause some independent
Speculation over the possibility of cause and effect.
Early History -13-
This question was So important that it was brought by Suits to the
attention of Vice President R. E. Luebbe, general counsel of the Company.
It was recognized that the possibility of liability for damage from cloud-
seeding experiments was a very worrisome hazard in this new form of
cloud experimentation. Since such a threat to the share owners’ money
would not be balanced by any known gain to the Company’s products or
business, there was great reluctance to incur risks of uncertain but
potentially great magnitude.
It was considered particularly important for this reason that any
seeding experiments be conducted under government sponsorship. No
further seeding flights were made until such sponsorship was provided.
A contract (W-36-039-sc-32427 req. EDG 21190) was finally re-
ceived from the Signal Corps covering ‘‘research study of cloud particles
and cloud modifications’’ beginning February 28, 1947. It covered cloud
modification by seeding, plus investigations of liquid water content, par-
ticle size, particle distribution, and “‘vertical rise of the cloud in respect
to the base.”’
An important part of the contract was a subparagraph stating that
‘the entire flight program shall be conducted by the government, using
exclusively government personnel and equipment, and shall be under the
exclusive direction and control of such government personnel.’’ The Re-
search Laboratory immediately notified all those involved in the research
‘‘that it is essential that all of the G.E. employees who are working on this
project refrain from asserting any control or direction over the flight pro-
gram. The G. E. Research Laboratory responsibility is confined strictly
to laboratory work and reports.”’
Although the contract was a Signal Corps contract, the project actu-
ally had joint sponsorship by the U.S. Army Signal Corps and the Office of
Naval Research, with the close-cooperation of the U. S. Air Force, which
furnished airplanes and the associated personnel.
The title of Project Cirrus was not applied immediately. It went into
effect officially on August 25 of that year.
Ill - GETTING ORGANIZED
CONTRACTUAL HISTORY
The work done on Project Cirrus and the activities leading up to
it were covered by several contracts with the government.
The two research projects, involving first the work on gas masks
and smoke filters and then the work on smoke generators, extended over
a period from October 1940 through February 1944. This work was done
under two contracts (NDCre-104 and OEMsr-131) with the Office of
Scientific Research and Development.
From October 1943 through June 1946, precipitation static research
was carried on under Signal Corps contract W33-106-sc-65 and, subse-
quently, under Air Force contracts W33-038-AC-9151 and W33-038-AC -15801.
The meteorological research. which became Project Cirrus, was
supported for a time by the General Electric Company. In February 1947,
the first of three Signal Corps contracts (W36-039-sc -32427, W36-039-sc-
38141, and DA36-039-sc-15345) was signed. The last of these remained
in force until the end of September 1952.
ORGANIZATION
The over-all direction of the project and the formation of broad
matters of policy were entrusted to a Steering Committee, consisting
of representatives of the three military branches of the government co-
operating in the project. Dr. Irving Langmuir and Dr. Vincent J. Schaefer
of the Research Laboratory served as consultants on the committee. The
military personnel was as follows:
Signal Corps. Dr. Michael J. Ference, Jr., chief, meteor-
ological branch, Evans Signal Laboratory,
Belmar, N. J. His alternate was Dr. C. J.
Brasefield of the same unit of Belmar.
Navy. E.G. Droessler, geophysical branch, Office
of Naval Res., Navy Department, Washington.
His alternate was Commander R. A. Chandler.
Getting Organized =16<
Droessler was succeeded in the summer of 1950
by Lt. Max A. Eaton. Commander Chandler was
succeeded in the summer of 1949 by Commander
G.D. Good, DENO©. (Air).
Air Forces. Major P. J. Keating, chief, Weather Equipment
Flight Test Facility, Middletown, Pa. Major
Keating was succeeded 3/23/49 by Col. N.C.
Spender of the Air Weather Office, Washington.
Major Keating had no alternate; Col. Spender’s
alternate was Lt. Col. J. Tucker of the Elec-
tronics & Atmospheric Branch at Washington.
The activities of Dr. Langmuir, Dr. Schaefer, Dr. Vonnegut, and others
of the General Electric Company’s Research Laboratory staff were limited
by the Steering Committee to laboratory work and analysis. The General
Electric scientific group came to be known to the personnel of the project
as the Research Group. In addition to Langmuir, Schaefer, and Vonnegut,
this group included Messrs. Kiah Maynard, R. E. Falconer, Raymond Neu-
bauer, Robert Smih-Johannsen, Duncan Blanchard, George Blair, Myer
Geller, Victor Fraenckel, and Charles Woodman.
An Operations Group was established by the Steering Committee early
in the life of the project to plan, co-ordinate, and control all project air oper-
ations, aSsist in the assembly and analysis of all technical data obtained,
provide all necessary meteorological information and service required for
the efficient conduct of the project, and take whatever action was necessary
to fulfill those requirements. This group would contain all military and
civilian personnel necessary to fulfill those functions, and it would be under
the direction of an Operations Committee. This committee was set up to
‘‘assume full responsibility for, and, therefore, exercise complete freedom
of action in the initiation of plans for, and the control of, all project air oper-
ations to be conducted in the vicinity of Schenectady.’’
The Operations Committee was set up, like the Steering Committee, to
include representatives of the three services, plus Kiah Maynard of the Re-
search Laboratory as General Electric consultant. It went through numerous
changes of personnel. The initial membership, and subsequent changes, were
as follows:
1. Lt. Comm. Daniel F. Rex, USN, chairman; Capt. C. N.
Chamberlain, USAF; Roger Wight, Signal Corps; Mr.
Maynard.
2. Wight was succeeded by Samuel Stine in August, 1947.
Getting Organized -17-
3. In June, 1948, Mr. Stine became chairman and Lt. Comm.
ih, B, Faustwexecutive officer.
4, In the fall of 1948 Major Rudloph C. Koerner, Jr. became
chairman, Rex and Stine left the committee, and Capt. J. A.
Plummer, USAF, was added.
5. In February, 1949, Lt. Comm. Paul J. Siegel became ex-
ecutive officer and Lt. Comm. Faust, operations officer.
6. In April, 1949, Faust was succeeded by Capt. Carl F. Wood
as operations officer, Faust becoming data control officer.
Plummer left the committee. Membership from then on:
Koerner, Siegel, Wood, Faust, Maynard.
The initial personnel of the operations group consisted of six repre-
sentatives of the Signal Corps, six of the Air Force, and six of the Navy.
Although the number of General Electric people working on the project re-
mained fairly constant at a figure of six or seven, the government repre-
sentatives varied widely in number. As a consequence, the total personnel
of the project varied also, running as high as 40 or 41 persons at various
times when activities were at their peak. These included crewmen for the
planes, weather technicians, and civilian employees for such services as
photography. A total of 33 persons went on the Puerto Rico operation, and
37 went on the second trip to New Mexico.
An alphabetical list of the members of Project Cirrus at one time or
another is attached as Appendix |.
FLIGHT PROGRAM
At the outset, and until June 1, 1947, Project Cirrus test flights were
made by a plane from the Weather Squadron assigned to the Signal Corp.
This plane visited Schenectady six times, and a total of five seeding flights
were made. Olmsted Field at Middletown, Pennsylvania, was the base of
operations.
It was Soon discovered, however, that many delays in carrying out
flights could be traced to this geographic separation of the Operations
and Research groups. Accordingly, in the summer of 1947, all flight
operations were transferred to Schenectady. Headquarters for the Opera-
tions Group was established at the General Electric hangar at the Schen-
ectady County Airport.
The facilities steadily expanded until, at the end of 1948, they con-
sisted of a total of 1830 square feet of office, operations, and storage
Getting Organized -18-
space, including a flight tower, weather office, administration office, dark
room, navy cage, Recordak room, operations office, analysis room, and
a parachute-and-stock room. In addition to this, about 640 square feet of
conference room was available whenever required. In the same category
was a room in the hangar for aircraft, when a heated area was needed for
installation work, repairs, or other reasons.
On call were two aircraft mechanics, two shop men, two transcribers,
and an instrument man. A full-time secretary handled reports, correspon-
dence, telephones, etc.
To facilitate flight operations, two Weather Bureau teletype circuits
were installed, as well as a Teletalk system connecting all offices. This
could also operate a public-address system in the hangar and the ramp. In
addition, connections were made through two leased wires to the Boston CAA
control center and the Army Airways control center at Middletown, Pa.
At the hangar, a repair station was available. Guards were assigned
for the protection of aircraft and equipment, and standard aircraft fire-
fighting equipment with trained personnel was on hand for emergencies.
At first the number of aircraft assigned to the project was disappoint-
ingly meager, but eventually this situation was corrected. At one time as
many as six planes were available--three from the army and three from
the navy.
Active flight operations ran from the establishment of the project in
March, 1947, until August, 1950, when the Operations Group was disbanded
at the suggestion of the Research Group. (This move was made in the inter-
ests of economy, for most of the objectives of the flight program had by that
time been accomplished.)
A list of all the flights made by Project Cirrus is attached as Appendix
II,. This list includes the flights made in rented planes before the establish-
ment of the project. It also includes the flight numbers for the time after a
system of numbering was instituted.
Although a brief statement of the location and purpose of each flight
is also given in Appendix Tl, this information is not supplied in detail. It is,
rather, summed up in connection with the discussions which follow of the
individual studies and operations. Detailed descriptions of the flights are
available in flight folders located, at the time of this writing, in the files
of the Weather Station in the Laboratory penthouse.
Getting Organized -19-
GROUND OPERATIONS
In addition to the flight program, the Operations Group had the re-
sponsibility for condicting numerous operations on the ground, These
operations were of two kinds: photography and silver iodide seeding.
When it became apparent that such operations would be necessary as
part of the project from time to time, a system of numbering each operation
was established. A record of all the numbered ground operations was main-
tained by the Operations Group, and a tabulation from this record is attached
as Appendix III.
WEATHER STATION
Weather observation being essential to operations of the type carried
on by Project Cirrus, one of the first steps to be taken by the Operations
Group was to set up a complete weather-observing station as part of the fa-
cilities at the General Electric hangar. Daily radio contact was established
with the Weather Equipment Flight Test Facility at Middletown, Forney ivenla,
and circuits for weather teletype services were installed.
The primary requirements of the weather station were agreed to be as
follows:
1. Preparation of aerological flight data prior to take-off
on flight tests.
2. Gathering of data to supplement that obtained in the air
on seeding missions, gathered after the flight for the area
concerned during the time of test.
3. Co-operating with the Research Group in its study of weather-
analyzing instruments and test flights, and supplying it with
such special weather reports as needed for analyzing purposes.
In order to meet these requirements, the Weather Station performed
the following functions:
1. Daily small-cloud maps were prepared of conditions dur-
ing the last hour before take-off on test flights, covering
an area having a radius of 200 miles from the Schenectady
County Airport.
2. Daily flights were made to record the air conditions up
to 8000 feet above the airport.
Oo eae
Getting Organized -20-
3. Radiosonde data above freezing level were obtained daily
from Albany.
4, Daily surface weather maps were prepared of the com-
plete Eastern United States area.
5. Data were obtained daily of the winds aloft for the Eastern
United States.
6. Local actual weather observations were made hourly.
7, After each test flight, cross-sections of the areas seeded
were prepared, based on reports of flight personnel and
teletype weather reports.
When the Operations Group was disbanded in 1950 and the facilities at
the General Electric Hangar were abandoned, the Weather Station was trans-
ferred to the penthouse of the Research Laboratory at the Knolls.
Through the Office of Naval Research, two navy men had a lengthy
assignment to the project as aerologists, and as such they contributed much
valuable assistance to the study of general and specific problems encountered
in the various research studies. These men were Lt. (jg) W. E. Hubert and
H, J. Wells, AGC. (Lt. Hubert was succeeded in 1951 by Lt. Cdr. C. E. Tilden.)
A partial list of studies made by these men is included on pages i and ii of the
ano ewe yoo the final report on Contract W-36-039-sc-38141 dated July
SOs 1951)
PHOTOGRAPHY
Another very important activity essential to the success of the project
was photography of various kinds, From the outset it was found that complete
evaluation of the results of the various seeding experiments could not be
made without taking pictures.
Both still and motion-picture types of photography were used, In addi-
tion, special techniques were adopted. For example, by means of lapse-time
photographs it was possible to speed up movies in order to obtain a better
grasp of the changes taking place in a cloud. Also, by the use of stereoscopic
equipment, it was possible to produce three-dimensional views.
A photographic darkroom was provided as part of the Ground Operations
facilities at the General Electric hangar. When the Operations Group was dis-
banded in 1950, darkroom facilities were provided in the penthouse weather
station at the Knolls.
Getting Organized -21-
So important was photography considered in the active phase of the
project, when the Operations Group was functioning and regular test flights
were being conducted, that many civilian professional photographers were
employed in addition to those provided by the Signal Corps. On the second
New Mexico test operation, six photographers made the trip from Schenectady
to Albuquerque. During the Puerto Rican test operation, over 100,000
frames of lapse-time pictures were taken in color. The load on the darkroom
at the General Electric hangar in Schenectady became So great that a photo-
graphic trailer was obtained from the Signal Corps Engineering Laboratories
to relieve the congestion.
One print of each photograph was, at the time of the preparation of
this report, on file in the Knolls penthouse weather station, plus virtually
all motion pictures (some are in the possession of Schaefer). All negatives
are filed in the photographic vaults of the Signal Corps Laboratory at
Belmar, New Jersey.
INSTRUMENTATION
A considerable portion of the time and activity of Project Cirrus per-
sonnel was spent on the development of special instruments, tools, and equip-
ment essential to the project. As in any new undertaking in which there is
little or no previous experience, many new devices of this type had to be
designed, or old ones had to be adapted to special requirements. In addi-
tion to Schaefer’s simple cold chamber, which became a standard item of
meteorological research in the field of cloud physics, the more important
equipment of this type follows:
Dry Ice Dispenser. One of the first instruments which had to be de-
veloped was an automatic dry ice dispenser. 79) This was devised
(Schaefer -Falconer-Kearsley) for use in an airplane, to allow a continuous
release of dry-ice pellets during seeding operations.
Dry Ice Crusher. This was a device (Schaefer-Falconer -Kearsley)
for reducing blocks of dry ice to usable fragments for seeding purposes.
It greatly reduced the time required for preparing this material for a
seéding run.
76)
Silver Iodide Generators. A number of different methods for the gen-
eration of silver-iodide smokes were studied by Vonnegut early in the his-
tory ve ike project. One method vaporized silver iodide from a het fila =1>n
ment. 3 Another involved the use of a small electric furnace. o) A
third method vaporized silver iodide from a string in a flame and then
caused a very fine smoke by rapidly quenching the flame with a blast of
compressed air. 56) A fourth introduced silver iodide into flares of the
Getting Organized -22-
standard fireworks type.(57) A fifth technique produced silver-iodide smokes
by first producing a silver smoke with an electric arc and then converting
the silver. pe panucks to silver iodide by the addition of iodine vapor to the
smoke.
In addition to these, two other techniques were devised which were well
suited to large-scale seeding. In one, a solid fuel, such as charcoal, impreg-
nated with a silver-iodide solution, was burned.\°7,68) The silver iodide
vaporized and then condensed in the form of a fine smoke. In the other tech-
nique, a Solution of silver iodide and acetone was atomized in a spray nozzle
and burned, vaporizing the silver iodide. 73) The silver-iodide vapor
rapidly condensed when it mixed with the cool air of the atmosphere, to form
a smoke of very small particles, the size of which could be varied over a wide
range. A later design of this generator, adapted for use in flight, was found
to be simple and reliable.
Camera Clinometer. It became evident in early flights that it would be
necessary, when photographing seeded areas, to know the vertical angle at
which the camera was pointed. A very simple device was made (Langmuir-
Falconer) to attach to the camera to indicate this angle. i
Flight Instruments. Standard instruments often had to be modified,
and new ones were occasionally developed. For example, a device was
evolved (Maynard-Falconer) to record the movement of the airplane ‘‘stick’’
for correlation and measurement of vertical acceleration.
‘‘Weather’’ Instruments. But it was in the field of weather observation
and atmosphere studies that most of the instrument development occurred. Some
of the early devices were special rods (Falconer-Maynard) to be mounted on
the airplanes to determine the rate of icing; 75) an air decelerator (Schaefer-
Falconer) to assist in sorting out rain, snow, dust, or cloud particles from the
eee as the plane passes through; 5 ) and a cloud- particle gun
(Se tetas yeteon) for sampling the cloud-droplet size distribution in
clouds. An attempt was made to develop a cloud-particle ranging instru-
ment for airplane use to provide a continuous record of the distribution of
particle sizes in a cloud, but without success.
Cloud Meter. An important early development was a cloud meter
(Schaefer-Falconer), designed to provide information which would give a
eS Ue oi; he th Boy ae effective particle sizes in the various portions of a
cloud. This device, embodying a continuously moving tape im-
pregnated with a water-sensitive dye, gave a satisfactory indication of the
amount of cloud particles collected.
Condensation Nuclei Detector. Another important instrument (Vonnegut)
was one for obtaining a continuous record of the concentration of condensation
nuclei in a given air sample. 67) This involved a Simple adaptation of the
Getting Organized -23-
cloud-chamber technique. Also a very simple pocket-size unit was devised
for making spot checks of the relative numbers of such nuclei in a given
sample.
Vortex Thermometer. A development of much significance was the
design by Vonnegut of an instrument, the oe thermometer, for use by
airplanes in measuring true air temperature. 66) The usual type of ther-
mometer is unsatisfactory for this purpose because of aerodynamic heat-
ing caused by the rapid movement of the airplane through the air. The
vortex thermometer reduced these aerodynamic effects to a negligible
amount. Also, for the first time, it made it possible to give a quite accu-
rate measurement of the temperature in a cloud. Furthermore, an indica-
tion of true air speed can be provided by measuring the difference in
readings given by a vortex thermometer and one exposed in the normal
manner, because the deviation from true temperature of a normal ther-
mometer varies with the speed of the plane. But it was found that the
vortex whistle (see below) showed greater possibilities for this application.
Vortex Speed Indicator. An outgrowth of the development of the
vortex thermometer was the adaption of the principles involved to the
production of a musical note (Vonnegut). As the pitch of the note produced
in such a manner varies with pressure, such a whistle could be used as
the basis for measurement of true air speed and air mileage of airplanes, (71)
Rain Catcher. A tool found very useful in rain studies aloft was a
rain catcher, developed (Langmuir-Schaefer-Maynard) to give the average
value of the precipitation in the air for approximately each thousand feet
of flight. ‘The device involves the use ofa rain scoop, a tube whose exit
velocity can be controlled, and a group of Storage containers. 82
Portable Cold Chamber. A simple but effective cold chamber was
designed by Schaefer, which could be carried about for field studies. It
consisted of a small rectangular wooden box lined with copper sheeting
and having a copper inner chamber. A charge of five pounds of crushed
dry ay was found to hold the temperature below -10°C for three
hours.\83,84,86)
Ice Nuclei Detectors. Since one of the important properties of the
atmosphere as related to the persistence of supercooled clouds is the
presence of ice-forming nuclei, considerable effort was expended toward
the development of an instrument which would provide a continuous, auto-
matic record of the quantity of such nuclei in the air at any given time.
Two developmental instruments were devised, but difficulties were expe-
rienced with both of them, and neither was brought to a satisfactory
degree of perfection. One device (Schaefer) made use of the tendency
Getting Organized -24-
of a thin water-soluble film of polyvinyl alcohol to supercool. (41) The other
(Vonnegut) utilized the cooling effect of the ice crystals when they struck a
hot wire carrying an electric current.
Uniform Particle Generator. A useful tool in the study of cloud physics
is an apparatus for producing particles of uniform size, developed (Vonnegut)
during the work on one of the ice nuclei detectors. 70) With it, particles were
produced in sizes down to about 10 microns diameter.
Salt Particle Detector. An apparatus was constructed (Vonnegut-
Neubauer) that detects and counts aerosol particles, such as salt particles,
by the pulses of light they produce when they enter a hydrogen flame. Ob-
servations showed that the concentration of large sodium-containing particles
in the atmosphere is subject to considerable fluctuation,(74A
Cloud Chamber. A very simple but effective adaptation of the con-
tinuous eu i ee was developed by Schaefer, using water instead of
aleohols* =? It gave promise of considerable value in conducting quanti-
tative experiments with a controlled atmosphere.
Aerosol Precipitator. A very simple apparatus was constructed by
Vonnegut to precipitate aerosol particles from the atmosphere on a strip
of paper. It was found useful in the study of condensation nuclei in the at-
mosphere,
Snowflake Recorder. This device was developed (Schaefer-Falconer-
Kearsley) to record the type and concentration of snow crystals reaching the
ground during the storm period of the winter season, It utilizeda strip of
paper on which was rubbed a water-sensitive dye. 78
Cloud Type Indicator (Schaefer-Falconer), By measuring the daylight
from a small portion of the northern sky, it was found that the variations in
reflection caused by blue sky or various cloud types which passed this area
Bose at a curve which could be interpreted in terms of particular types of
cloud.
IV - LABORATORY STUDIES
The interest and activity in cloud seeding and the fundamental physics
of clouds, following the initial experiments, were so varied that it is diffi-
cult to.give an orderly account of the progress in this field, Research both
in the laboratory and in the atmosphere continued to reveal new and inter -
esting facts. The contents of this section of the history consist of summaries
of the more important laboratory studies in this field which were conducted
by the Research Group of Project Cirrus.
PERSONNEL
It would be difficult, if not impossible, to list the names of all the
people contributing to the laboratory studies of the project. But twelve
persons should be mentioned who took part, either continuously through-
out the life of the project, or at one time or another during its existence.
Dr. Irving Langmuir, under whose direction the project evelved,
planned the methods and techniques for the various programs, analyzed
flight results, and set up procedures for the routine analysis of such re-
sults. He also reduced to convincing mathematics many of the theories
evolved.
Dr. Vincent J. Schaefer, who worked with Langmuir in the planning
of the project, carried out both field and laboratory experiments on the
fundamental processes involved in changes of cloud forms,
Dr. Bernard Vonnegut also carried out extensive field and labora-
tory experiments on subjects associated with the project. Particularly
he concentrated on theories and techniques associated with the use of
silver iodide for seeding.
Raymond E. Falconer worked on various phases of instrumentation
of the flight planes, on laboratory studies, and on other related problems.
He worked closely with Langmuir in his periodicity studies. After the
termination of the Operations Group, the establishment and maintenance
of a weather station in the Knolls penthouse was his primary responsibility.
Victor Fraenckel served as General Electric representative on the
Steering Committee and as contract liaison.
Kiah Maynard was the Research Laboratory representative on all
flight tests and on the Operations Group when it was active. He gathered
data and maintained records of all flight tests. He was associated with
Falconer in the operation of the weather station at the Knolls penthouse.
Laboratory Studies -26-
Raymond L. Neubauer was associated with the later stages of the project
in the development of instruments and studies of silver-iodide smokes.
Robert Smith-Johannsen, associated with the project during its earlier
history, was principally concerned with the study of the supercooling of water.
Duncan Blanchard was temporarily associated with the project in con-
nection with the study of water droplets.
Myer Geller, temporarily associated with the project, contributed im-
portant calculating work.
Charles Woodman, temporarily associated with the project, contributed
important mathematical work.
Arthur Parr, a Research Laboratory machinist, built almost all the
special equipment and developmental instruments involved.
ICE NUCLEI
One of the most important phenomena associated with the study of the
physics of clouds is the formation, distribution, and relative abundance of
nuclei for the formation of ice crystals. This subject, therefore, occupied
the attention of the principal members of the Research Group to a greater
or less extent throughout its history.
Considerable work was done in developing instruments and methods
for detecting the presence of, and counting, such nuclei in the atmosphere.
Relatively early in the history of the project, a station was established by
Schaefer at the Mt. Washington Observatory for regular observations of the
concentration of such ice-forming nuclei, and these observations continued
over five years. Subsequently, Schaefer found in the laboratory that certain
kinds of soils, when dispersed a a dust, were moderately good nuclei under
certain atmospheric conditions. 43
At the time of writing this report, the number of ice nuclei needed in
a supercooled cloud to initiate a chain reaction (see page 28) was not yet
known, but evidence found early in the history of the project, suggesting
that a critical concentration is found in the range of 10,000 to 50,000 nuclei
per cubic meter, has consistently been strengthened since. o4
Observations of ice nuclei were also conducted at the Research and
Development Division of the New Mexico School of Mines at Socorro, with
whom the scientists of Project Cirrus maintained a close liaison.
Laboratory Studies -27-
A significant fact resulting from the Mt. Washington studies was
the rarity of Sal high concentrations of active ice-forming nuclei
in the atmosphere. If the observed results are a true representation
of the average mean condition of the atmosphere, it is obvious that, by
the artificial introduction of sublimation nuclei into the atmosphere, man
possesses a powerful method of modifying many cloud systems.
One prolific source of ice-forming nuclei might be the Great Plains
and the more arid regions immediately adjacent to the Continental Divide.
Wind storms, dust devils, and strong convective ey could easily ac-
count for the formation of ice-forming nuclei aerosols. 47)
It Seems probable ey ie smoke produced by forest fires is a
poor source of such nuclei. 47) An attempt was made to determine the
role that bacteria and the spores of fungi might play in this respect (17)
and to evaluate the role of industrial smokes of various’ kinds. 59)
Adiabatic Expansion of Gas. An important contribution to the early
knowledge of meteorological phenomena was made through Vonnegut’s
observations that, when gas is cooled to below -39°C by adiabatic expan-
Sion, very large numbers of ice crystals are formed.(60) For example,
the low temperature praduced at airplane propeller tips and wings can
seed supersaturated air or Supercooled clouds, resulting i persistent
vapor trails or cloud modification. Cwilong had reported\® that ice
erystals could be produced by this method, but he apparently had not
appreciated the enormous numbers which are so produced.
It was found that the adiabatic expansion resulting from the bursting
of a rubber balloon a millimeter in diameter produced over 10,000,000
ice crystals. Schaefer made a popgun which did the same thing, lending
itself to careful control of temperature, pressure, and humidity.
This provided corroboration of conclusions already reached with
dry ice and furnished additional quantitative data which were found very
useful.
Chemical Effects. An interesting effect noticed by Vonnegut while
carrying out Some Studies of ice crystals in a cold chamber was that
the presence of normal butyl alcohol caused the crystals to form as hex-
agonal columns instead of hexagonal plates, (08) The phenomenon was
studied by Schaefer in some detail, but no practical application of the
findings was developed.
Laboratory Studies -28-
Spontaneous Formation. Work done by Schaefer and others as early
as 1946 indicated that ice crystals formed spontaneously in water-saturated
air when the temperature reached the neighborhood of -35 or -400C. Schaefer
conducted quite a bit of research into this subject of spontaneous forme
and determined that the critical temperature was -38.9+ 0.1 degrees. o4
This phenomenon is probably of considerable significance in relation
to the formation of cirrus clouds and ice crystal fogs in the free atmosphere.
Structure. Schaefer’s study of the various types of snow crystals, which
started before the establishment of Project Cirrus, continued throughout the
project. In 1948 he published a simple yet seat list of ten types of solid
precipitation for classification purposes, (39 In slightly modified form this
classification is now in use throughout the world. ‘
Crystal Growth and Multiplication. Experiments made by Schaefer in
1949 indicated that snow particles tend to shed minute fragments of ice when
they are placed in air slightly warmer than their own temperature. An ice-
forming nucleus appearing in a supercooled cloud grows rapidly, especially
in the temperature range of -12 to -16°C, where the difference between the
partial vapor pressure of ice and of water passes through a maximum. When
the crystal becomes large enough, it sheds a considerable number of ice parti-
cles as it falls through the cloud. These particles then serve as new nuclei
and repeat the cycle. In this manner, a few ice-forming nuclei in a cubic
meter of cloud may start a chain reaction which, within_a few minutes, could
shift a supercooled cloud to a mass of snow crystals.
A laboratory study was made to determine the factors of importance
for obtaining the maximum rate of snow crystal growth.
SILVER IODIDE
After the discovery that silver-iodide smokes serve as an excellent
nucleus for the formation of ice crystals, the project was faced with the prob-
lem of finding some way of generating the smoke efficiently and in quantity.
It was found that smokes consisting of exceedingly fine particles could be
easily produced by vaporizing silver iodide at a high temperature and then
rapidly quenching the vapor. This was readily accomplished by burning
Silver -iodide -impregnated charcoal or injecting a spray of silver-iodide
Solution into a hot flame. Simple generators based on this principle were
made which could produce 10 ?*nuclei per second--enough to seed from
1000 to 10,000 cubic miles of air per hour (65
Laboratory Studies -29-
A very interesting fact discovered as the result of one of Vonnegut’s
studies is that silver-iodide particles do not react immediately as ice-
forming nuclei when introduced into a supercooled cloud of water droplets.
Even 50 minutes after introducing a smoke sample into the cold chamber,
ice crystals could be seen to form at a measurable rate. The general
conclusion reached as a result of this study wee iat the rate of reaction
at -18°C is 30 to 40 times faster than at 10°C, ©
The first unambiguous results in cloud seeding using silver -iodide
generators were obtained in 1948. Silver-iodide nuclei produced by one
of Vonnegut’s generators installed in an airplans resulted in cloud mod-
ification similar to that produced by dry ice.
Experiments were conducted to determine whether the burning of
charcoal particles used in silver-iodide seeding from an airplane would
be seriously affected by the moisture in clouds. It was concluded that
the bumping is not seriously affected if the charcoal is thoroughly ig-
nited.\°/
Some experiments were conducted to discover the value of a turbo-
jet burner as a Silver-iodide smoke generator. It was decided that such
a method might &S of value if larger generators were needed than those
already in use| 8)
Experiments were also made in tracing silver-iodide smokes after
their release by seeding generators. 6
The nature of silver iodide is such as to suggest the possibility
that its effectiveness as a seeding agent might be reduced by the action
of ultraviolet and near-ultraviolet radiation from the sun. Accordingly,
an investigation was made to determine its rate of decay under expected
conditions of radiation in the free atmosphere. The results of work in
this field not only by Project Cirrus, but also the New Mexico School
of Mining and Technology, suggested that far greater quantities of silver-
iodide particles might be required for seeding operations under conditions
of bright sunlight than would be needed at night or under conditions of
Cloud cover. But later work and observations indicatsd that the effect
of sunlight might not be as bad as was forecast. ol,72
Experimental work showed that it is possible to convert super-
cooled ground fogs to ice crystals by releasing silver-iodide smokes.
(56)
Laboratory Studies -30-
RAINDROP STUDIES
Although many of the details are still lacking, studies conducted by
Project Cirrus began to provide answers to the question of how rain is
formed.
In 1947, when reports were received of successful results obtained
by dry-ice seeding of cumulus clouds over Hawaii having a temperature
above the freezing point, Langmuir restudied theoretical calculations he
had prepared in 1944 in studies relating to work at Mt. Washington Ob-
servatory. As a result ae developed a theory which agreed very well with
the reactions reported, (/ )
According to Langmuir’s theory, actively growing cumulus clouds
having an average drop size of 20 microns, a liquid water content exceeding
2.5 G/M, and a vertical thickness of more than a mile are in favorable
state for starting a chain reaction. This could be achieved by introducing
water drops greater than 50 microns in diameter into the actively growing
part of the cloud.
Large drops in such a cloud would fall at a greater velocity than
would small drops. In falling, they would overtake and collide with the
small drops and thereby increase in size. In time the large drops would
become so large that surface tension could no longer hold them together,
and they would break up into two or more smaller drops. These in turn would
grow and break up, and the number of large drops would increase in this
manner by a chain reaction.
The process would not be sufficient to produce large numbers of
raindrops in a cloud without a vertical updraft. However, in the case of
clouds with suitable updraft conditions, many stages of the chain reaction
are carried out, resulting in the production of rain.
This chain-reaction theory led Langmuir to postulate that cumulus
clouds having sufficient updrafts could be seeded with a few large water
drops.
To determine the validity of several of the important phenomena in-
volved in this theory, various studies were initiated in the laboratory and
experiments conducted in the field. Blanchard devised a splendid method
for studying the properties of free-falling water droplets in air, using a
vertical wind tunnel. A series of striking stroboscopic photographs was
made, showing the oscillations, gyrations, Sree UE and fractures that go
on as water drops fall at their terminal velocity.
Laboratory Studies -31-
Another activity concerned itself with devising means of sampling
raindrops and measuring diameter. 3
| Seeding with water drops was carried out with apparent success in
tropical clouds. 41) This is more fully discussed in a later section of
this report.(Page 48)
CONDENSATION NUCLEI
Condensatién nuclei played an important role in the behavior of the
atmosphere. In 1948 Vonnegut devised a method of obtaining a continuous
record of the concentration of condensation nuclei in the atmosphere, (©7
Various experiments were conducted with this equipment, both aground
and aloft. The results suggest that the continuous measurement of the con-
centration of condensation nuclei may be very useful in meteorological in-
vestigations.
ELECTRICAL PHENOMENA
It was observed in 1943 by Schaefer that interesting atmospheric
electrical measurements could be obtained by connecting one end of a
shielded cable to an insulated needle presented to the sky and the other
end to a Suave recording microammeter, one side of which was well
grounded, Among the interesting observations made during successive
years was one to the effect that the data obtained with this instrument in-
dicated the passage of charged clouds over the observation point.
Continuous records were kept by Falconer from 1948 on, using the
data provided by this equipment, and an attempt was made to correlate
the measured corona-discharge currents with other meteorological phe-
nomena, such as frontal passages, wind direction, precipitation, and re-
flected light from the northern sky. It was found that there was generally
good agreement between such findings and those of other investigators.
Best correlations obtained with this equipment seemed to be with
frontal passages associated with the arrival of new air masses and the
occurrence of precipitation not necessarily local but possibly extending
to a radius of a few hundred miles. But correlation was also obtained
with wind shifts and pressure changes, Since frontal passages were
associated with those phenomena. There also seemed to be some re-
lation between certain instrument indications and small, sharp changes
in the ne een light from the northern sky, particularly in apparently
clear skies |
Laboratory Studies -32-
Workman-Reynolds Effect. When Workman and Reynolds announced
in 1948 their discovery of the formation of a large electrical potential when
water containing small quantities of certain salts is in the process of
freezing, Schaefer decided to check the experiments by an independent in-
vestigation. Accordingly, test equipment was set up and observations were
made.
The Workman-Reynolds electrical effects were immediately observed.
The results of this experiment have very important implications with re-
spect to the development of lightning in thunderstorms.
Electrical Atomization. Some qualitative experiments were made by
Vonnegut and Sie ana ig determine the effects of high voltage on the form-
ation of water drops. 74B) tt was found that streams of highly electrified,
uniform droplets about 0.1 millimeter in diameter could be produced by
applying potentials of from 5 to 10 kilovolts, ac or de, to liquids in small
capillaries. Aerosols of uniform size and having a particle radius of a
micron or less could be formed if the capillary was positively charged and
if liquids having low electrical conductivity were used. Aerosols formed in
this way showed the colors of higher-order Tyndall spectra.
DO RUDYSOr CLOUD EY DES
In connection with an investigation of snowstorm intensities, Schaefer
started measuring variations in sky brightness using a light-sensitive instru-
ment. Falconer subsequently carried on the measurements in more detail.
It was discovered that the variations in the curve made by this instrument
were a rather good indicator of the type of cloud cover prevailing during a
day. There seemed to be a typical trace for each general cloud type.
Such an instrument might be useful in automatic weather stations, to
give some indication of sky conditions in remote locations.
Test installations were made by Falconer at various points aground
and aloft, and considerable data were gathered.
ANALYTICAL WORK
Of great significance, both in connection with activities of the Research
Group and with those of the Operations Group, was the analytical work per-
formed by Langmuir. It constituted one of the most important contributions
to the project.
Laboratory Studies -33-
From the outset he studied and analyzed the various test flights of
the Operations Group, and extensive reports were prepared analyzing cum-
ulus and stratus cloud seedings. His analysis of the cumulus seedings over
Hawaii and the chain-reaction theory of rainfall which resulted have al-
ready been mentioned. (Page 30)
Langmuir paid particular attention to the seeding operations carried
on in New Mexico, and to the possible effects of silver-iodide seeding, and
these activities are described more fully in a later section of this report.
(Page 47)
Such a large quantity of data was accumulated by flight, field and
laboratory activities during the more active period of the project, that the
Research Group finally suggested early in 1950 to the Technical Steering
Committee that flight operations be terminated at Schenectady in order
that the accumulated data might be evaluated and reports prepared on
the findings.
V - CIRRUS AND STRATUS STUDIES
CIRRUS CLOUDS
The significance of cirrus clouds and the role they play in various
weather phenomena were, of course, subjects of intense interest to Pro-
ject Cirrus. Various studies of and experiments with such cloud forms
were conducted, although more attention was paid to stratus and cumulus
clouds.
A regular daily observation program was begun in 1947 to explore
the possibility of inducing the development of cirrus-type clouds under
clear sky conditions. It was believed that supersaturation with respect
to ice probably occurs fairly frequently at temperatures warmer than
-39°C in air devoid of foreign-particle nuclei. Lacking such nuclei, a
considerable degree of supersaturation could develop, as is often shown
by the generation of so-called vapor trails behind high-flying aircraft.
- To explore these possibilities, Falconer initiated a project in which
balloons carrying dry ice in open-mesh bags were released on a daily
schedule and followed by theodolite. Many of these produced visible trails
of ice crystals, and in several instances the trails were quite notice-
Bic oo, 18)
Several seedings were also carried out from an airplane in clear
air, using both dry ice and silver iodide. In clear air supersaturated
with respect to ice, the seeding operation produced a cloud made of ice
erystals. The results of these operations indicated that, if the humidity
is low, even at temperatures below -39°C, appreciable supersaturations
with respect to ice can exist without the formation of ice crystals. Ice
crystals can then be created, however, by seeding with either dry ice or
silver iodide.\’3 :
Natural Formation. In six of the Project Cirrus test flights a con-
siderable effort was directed toward obtaining photographic evidence of
the appearance of the tops of cirrus clouds. It was found that, despite the
various irregularities seen from below, the top of such clouds is extremely
mart:
Most meteorologists and weather students agree that a cirrus cloud
formation is often associated with the overrunning of cold air by a warmer
tongue of moist air. Whenever the moisture conditions in the warm over-
riding air reach saturation with respect to water and the colder air below
has a temperature of -39°C or colder, ice crystals will form spontaneously
at the inversion interface. The number of primary crystals that form will
depend on the concentration of condensation nuclei and ice nuclei in the
moist air mass. The number and size of secondary crystals that form will
Cirrus and Stratus Studies Ge
probably be some multiple of the effective number of condensation nuclei.
Since these conditions for the ice-crystal formation are of a marginal na-
ture, the variability and often unique appearance of true and false cirrus
clouds may be closely related to these spontaneous crystal formation
phenomena.
Based on this reasoning, Schaefer concluded that it is likely that the
concentration of supercooled water droplets at the transition temperatyre
of -39°C is of primary importance in the formation of cirrus crystals. )
Langmuir, analyzing the behavior of cumulus clouds, described an
action which he called cirrus-pumping. This occurs when, with few or no
nuclei present, the cloud rises to great heights. If it rises to a height when
the temperature gets down to -39°C or thereabouts, minute ice crystals are
formed in great numbers, almost instantaneously. These come into contact
with the supercooled water droplets in the cloud and immediately cause them
to freeze. This, in turn, liberates a large amount of heat simultaneously over
the whole top of the cloud, and this upper part rises still further, forming a
cirrus crown shaped something like a pancake.
The pancake grows in dimension and gets thinner, and it sometimes drifts
gradually off to one side, so that it assumes the general appearance of an
anvil--a type of cloud characteristic of the tropics. One large cloud of this
type, said Langmuir, might sometimes produce cirrus clouds which would
spread over 10,000 square miles. Outside of the tropics, they may often occur
Sune ay summer in semi-arid regions such as New Mexico, Arizona, or
Idaho.
Height, Temperature, etc. Some observations were made by the project
of the height of cirrus clouds and their temperatures.
STRATUS CLOUDS
Much more attention was paid to stratus clouds. The flight test of
December 20, 1946, for example, was cen ucted when the sky was com-
pletely overcast, and it produced snow. 12) Im the flight test of March 6,
1947, now under the auspices of Project Cirrus, seeding was conducted
on stratus clouds. Looking down on the cloud, it was observed, first, that
a deep groove had been produced along the top of the seeded area, and snow
fell. Soon the sky cleared up in a spectacular fashion, so that there was a
cloudless area 20 miles long and 5 miles wide where the seeding had taken
ee ey there were no other breaks in the overcast in any direc-
tion, Further tests on stratus clouds produced similar results.
Cirrus and Stratus Studies =e
The conclusion was therefore reached in the earliest days of the pro-
ject that cloud seeding could produce holes in stratus clouds. Thus a plane
should be able to clear a hole for itself. The result would be not only to
increase visibility but also to eliminate icing conditions.
Langmuir made an exhaustive analysis of the photographic data ob-
tained on these early test flights, reaching some very ete conclu-
sions regarding the nature and behavior of stratus clouds. lg
It was soon found that a very useful technique in seeding stratus
clouds was to seed in patterns--L shapes, race-track shapes, Greek gam-
mas, etc. Thus it would be possible to watch for modification'of the clouds
following the same pattern. And invariably modification did occur, agreeing
with the pattern of the seeding. In many cases clear areas were produced
in the cloud deck.
Among the stratus cloud studies made by the project were:
(1) The effect of seeding supercooled stratus clouds with
various amounts of dry ice and silver iodide.
(2) The optimum quantity of seeding agent required to pro-
duce large cleared areas in an otherwise solid deck of
supercooled clouds.
VI - CUMULUS STUDIES
The most spectacular, fruitful, and controversial results produced
by the activities of the project were those produced as a result of the work
on cumulus clouds. This work, which started in the earliest days of the
project, continued throughout its duration and let into some very inter-
esting other activities.
Flight tests on November 23 and 29, 1946, were made on isolated
cumulus-type clouds. The whole of each cloud was changed into ice
within five eee after seeding, and snow began falling from the base
of the cloud.\7°) But it was realized that experiments with small cum-
ulus clouds were of little interest, for the effects lasted but a few minutes.
Other experiments were conducted with cumulus clouds in the early days
of the project and, although many of them were changed to snow, the re-
sults were of comparatively little interest.
By the summer of 1947, however, some spectacular results were
obtained with cumulus clouds, especially with thunderstorms. ‘These were
so impressive that it was decided to make some studies of cumulus clouds
and thunderstorms in New York State’s Sacandaga Reservoir territory, not
far from Schenectady.
This reservoir is situated just south of the southeast corner of the
Adirondack Mountains. Evidence pointed to the probability that this large
body of shallow water provides the moisture which feeds thunderstorms in
eastern New York State. It was believed that the unusual conditions there
could be used to observe the effect of seeding the intense thunderstorms
developed. Actually, however, no seeding was performed there, although
many photographs were taken and considerable time was spent in a study
of conditions in that area.
HONDURAS
In 1948 and 1949, Langmuir visited Honduras, Guatemala, and Costa
Rica to study tropical cloud formations, and particularly to learn what was
being done by Joe Silverthorne, a commercial cloud seeder, in seeding
clouds for the United Fruit Company. The work was being conducted for
the purpose of testing out the possibility of controlling rainfall, and partic-
ularly in the hope of stopping blow-downs that result from winds associated
with thunderstorms, which occasionally destroy large stands of fruit trees.
At Langmuir’s suggestion, Silverthorne tried out a number of ex-
periments early in 1949 and made many worthwhile observations. It was
Sometimes desired to produce rain, and sometimes it was desired to pre-
vent rain. On the one hand, by overseeding the top of a high cumulus
cloud, rain would be prevented. The top of the cloud would float off into
_a higher altitude, where it would be blown away by the counter trade wind.
Cumulus Studies -40-
If, on the other hand, the cloud was seeded just above the freezing level,
heavy rain might be produced. Similarly, water seeding by means of water -
filled balloons released from airplanes might dissipate a cloud and produce
rain at low altitudes, but it seemed that in such instances dry-ice seeding
would be much more effective.
April 18, 1949. The results of the flight on this day, with Langmuir
seco qpanying Silverthorne aloft, were so outstanding as to merit detailed
comment. The following is extracted from an account of the flight by
Langmuir in the Project Cirrus report to the government of July 30, 1951:
‘‘We flew up to Point Sal and found a mass of dry air above
the moist air coming from the sea at an altitude of about 6000
or 7000 feet....From a height of about 8000 feet, looking South,
a whole panorama of high cumulus clouds could be seen rising
above the smoke, which extended up to about 11,000 or 12,000
feet further inland, although it was much lower than this near
the sea.
*‘A large cloud was found which rose, I believe, to a height
of about 25,000 feet, and we seeded it by making a series of short
passes into the cloud at an altitude of approximately 21,000 feet--
two pellets* about one inch cubed being dropped into the cloud
at 50-second intervals during these passes. The whole circuit
of the cloud was made, and then the plane moved off a short
distance, enabling us to see the effect produced.
‘‘A band around the cloud, perhaps 500 or 1000 feet high,
was observed which obviously consisted of ice crystals and
which ultimately detached itself from the lower part of the
cloud and floated off as a huge mass of ice crystals that could
be seen for a long time.
‘‘After the top of this cloud had turned to ice crystals
and had detached itself, there was left under this cloud
nothing but a group of lower clouds that reached only about
14,000 feet, which-was below the freezing level. Later we
flew down among these clouds and found that cloud bases
had gone down from 12,000 feet to about 7,000 feet. It was
difficult to see whether any rain was falling because of the
smoke, but from the lowering of the cloud base we concluded
Se Nr eee ee en ee a ea ee ee ae
*Dry ice.
Cumulus Studies -41-
‘that rain had fallen from the lower part, while the top
of the cloud had detached itself and floated off towards
the northeast.
‘Shortly after seeding this cloud with 10 to 12 pellets,
we picked out a smaller cloud nearby whose top reached
about 20,000 feet and dropped one single pellet of dry ice
one inch cubed on this cloud. About 8 or 10 minutes later
we found that this whole cloud had changed to ice crystals.
We flew through the ice crystal cloud and verified the fact
that they were entirely ice crystals. You could see them
blowing into the cabin, and we also found that the cloud grad-
ually dissipated. It probably rained out from the lower part
of the cloud but this was down in the smoke level where we
could not see it, and the top of the cloud then gradually mixed
with the surrounding dry air which had been deprived of its
source of supply of moisture from below.
‘In other words, on this day we had beautiful examples of
two effects that can be produced by seeding with pellets of
dry ice. First the seeding of the top of the cloud can cause
the top to float off from the lower part. However, in this
ease some of the ice crystals reach the lower part of the cloud
and cause rain to dissipate it. In the other seeded cloud,
which was much lower and reached only a few thousand feet
above the freezing level, the whole cloud rapidly dissipated
as the upper part changed to ice and the lower part rained out.”’
The results of the flight of April 18 constituted for Langmuir a won-
derful demonstration of the effectiveness of single pellets of dry.ice for
modifying large cumulus clouds. Such single-pellet seeding had a number
of practical advantages.
It quickly became obvious to Langmuir that the set-up for carrying
out cloud-seeding experiments in Honduras was unique. Silverthorne made
flights virtually every day, and, somewhere within a 150-mile range, clouds
were nearly always found suitable for seeding. Such clouds were almost
always orographic and associated with certain mountains.
Many interesting experiments were conducted, and almost always
the clouds could be profoundly modified with single pellets of dry ice.
The latter part of Silverthorne’s seeding operations used 10-20 peNets,
presumably to make sure the crystals were more uniformly distributed.
Cumulus Studies -42-
PRIEST RIVER SLUDY
Meanwhile the study of cumulus clouds had been approached from an-
other angle. Early in 1948 a visit was paid to the Research Laboratory and
Project Cirrus by H. T. Gisborne of the Northern Rocky Mountain Forest
and Range Experiment Station, United States Forest Service. Gisborne was
in charge of fire research for Region No. 1. He wanted to learn more about
cloud modification studies.
This fitted in nicely with Schaefer’s interest in the same subject. He
was anxious to study thunderstorms in a good breeding ground, and Gisborne
wanted to see if anything could. be done to reduce forest fires by thunderstorm
modification.
As a result, Schaefer visited the Laboratory at Priest River, Idaho, in
July of that year (1948). He conducted quite a study of conditions there and
made rather complete recommendations for a plan of future activity--a plan
_ which should produce beneficial results from both Gsnepomts: Gisborne’s
practical aspects and Schaefer’s theoretical ones.
Actually, the recommendations were never put into effect. A consid-
erable force for the completion of the project disappeared with the death of
Gisborne. Although the project is still incomplete, interest still exists, how-
ever, both at Schenectady and at Priest River.
RESULTS IN HAWAII
Further data, supplied from still another source, had some unexpected
and very interesting implications and results.
Early in 1947 a request for information on techniques of dry-ice seeding
was received from the Pineapple Research Institute of Honolulu, Hawaii. This
information was supplied by the Research Group of Project Cirrus, which had
been supplying similar information to meet numerous requests Since the pub-
lished reports appeared of Schaefer’s historic snowmaking flight over Pitts -
field in 1946. But in this case there was an unexpected aftermath.
In October, Honolulu newspaper accounts were received in Schenectady,
describing experiments carried out over the island of Molokai by Dr. L. B.
Leopold and Maurice Halstead of the Pineapple Research Institute. A few
weeks later, copies of a preliminary report were received from these two
men, describing interesting results obtained by dumping dry ice into cumulus
clouds having temperatures above the freezing point.
Results in Hawaii -43 -
This was an important development. Although Langmuir had given
some thought to the effects of seeding nonsupercooled clouds, he hadn’t
done much about it, and this new work caused him to restudy theoretical
calculations bg had prepared in 1944 in connection with the work at Mt.
Washington. | )
He now had a new approach to the subject of weather modification:
the growth of rain.
RAIN CHAIN REACTION
The result was Langmuir’s chain-reaction theory of rain production,
in brief, as follows: A typical large drop of water grows in size as it falls
through the cloud, growing faster and faster until it gets so big that it
breaks up, producing smaller droplets. If there are rising air currents,
the little droplets will be borne aloft into the cloud again, growing in size
as they go, until they get so big that they start falling again. This process
continues in a chain reaction, causing the whole cloud to go over into heavy
rain. Under the right circumstances, according to this theory, seeding
with water would be just as good as with dry ice.
The outgrowth of this, in turn, was considerable work by Project
Cirrus to test Langmuir’s theory and apply some of its principles in prac-
tice. For example, to determine the validity of several of the important
phenomena which his theory postulated, laboratory studies were initiated
of the erpyye of water droplets and of the behavior of droplets floating in
the air.’’~’ These studies continued for a considerable period in the
laboratory, and some very interesting observations were made and data
collected. Later, the Research Group did considerable work in the stud
of the drop size and size distribution of various types of precipitation. 3
As another approach to the subject, an extensive series of exper-
iments was conducted to explore the possibility of inducing precipitation
or other modification in growing cumulus clouds by water seeding.
The complete exposition of the theory by Langmuir was a beautiful
example of theoretical analysis and mathematical calculation.(13) Among
other things, it reviewed the knowledge of cloud physics which had al-
ready been gained in the light of the new theory, summing up the probable
behavior of both stratus and cumulus clouds. It went so far as to suggest
that the chain reaction could, under the right conditions, be started by
introdicing even a single drop of water into a cloud, although the action
would be most rapid when many large drops were introduced near the
top of the cloud. It outlined the probable behavior of self-propagating
storms. It postulated that the phenomena that occur in artificial seeding
Cumulus Studies =44—
with dry ice or with water are essentially no different from those that occur
spontaneously in nature. ‘‘However,’’ it went on, ‘‘there will frequently be
cases where the cloud is not yet ready or ripe for spontaneous development
of snow or rain, although it may be possible to produce these effects by
seeding.’’ It concluded with the following significant summary:
‘‘When we realize that it is possible to produce self-
propagating rain or snow storms by artificial nucleation
and that similar effects can be produced spontaneously by
chain reactions that begin at particular but unpredictable
times and places, it becomes apparent that important
changes in the whole weather map can be brought about
by events which are not at present being considered by
meteorologists. I think we must recognize that it will
probably forever be impossible to forecast with any great
accuracy weather phenomena that may have beginnings
in such spontaneously generated chain reactions.’’
STUDIES IN PUERTO RICO
All these studies and tests which had been made, and theories which
had been evolved as a result, with regard to the nature, behavior, and modi-
fication of cumulus clouds were an important background to another signif-
icant milestone in the history of P pect Cirrus, That was the expedition
to Puerto Rico in February, 1949. 4
The objective of this trip was mainly to determine the type and physical
characteristics of the clouds that occur in Puerto Rico during the winter
months, particularly the month of February, and, if suitable clouds were en-
countered, to develop and possibly to evaluate water-seeding techniques. Con-
Siderable personnel took part in the project, a supply of planes was available,
and a large quantity of photographs was made.
At least two new precipitation sequences were observed, and considerable
data were accumulated to permit a better understanding of the processes in-
volved. Also studied was the trade wind inversion, a dominant feature which
controls cloud and precipitation development in the West Indies region during
February. A better understanding of this phenomenon should lead to a better
understanding of tropical meteorology.
The cumulus clouds were observed to have a different character than
those common in the eastern United States. Contacts made with interested
local people in Puerto Rico were expected to lead to the accumulation of
some excellent supplementary data on raindrop size, convergence of winds,
and the observation of double orographic cloud streams from the Liquillo
Mountains.
Cumulus Studies AG
The carrying out of successful ground-air operations on three dif-
ferent occasions, using lapse-time photographs as part of the ground
coverage, demonstrated conclusively to the members of the project the
value of carrying out such studies of clouds which develop in definite
cloud-breeding regions. Similar areas in the United States known to
possess such developments were Albuquerque, New Mexico, and Priest
River, Idaho. Schaefer had already visited Priest River, and arrange-
ments had been made for investigations and experiments there. And
a test mission had been conducted at Albuquerque the previous year,
details of which will be found in the next section of this report. (See
last paragraph on this page.)
Despite the fact that no suitable clouds were found for testing out
water-seeding techniques during the period, many valuable results were
obtained which it was expected would lead to a much better understanding
of the formation of rain in tropical clouds.
One of the very important results of the expedition was the obser-
vation of the important effect of salt nuclei on the formation of precipi-
tation in thin tropical clouds. Said one of the reports: ‘“‘This seems,
on first sight, to be of great importance in explaining the rain showers
which are of daily occurrence and random distribution in the vicinity of
Puerto Rico. Rarely is rain observed from such clouds in the eastern
United States.’’ Said Langmuir:
‘‘Observations in Puerto Rico in 1949 and in the Hawaiian
Islands in 1951 have shown that the rainfall depends on rela-
‘tively large particles of sea salt in the air, in accord with the
publications of A. H. Woodcock and Mary Gifford. Calcula-
tions of the rate of growth of salt particles indicate that it
should frequently be possible to induce heavy rainfall by
introducing salt into the trade wind at the rate of about one
tone per hour in the form of fine dust particles of about 25
microns in diameter. The heat generated by the condensation
may liberate So much heat as to produce profound changes
in the air flow and the synoptic conditions in neighboring
Pees
EARLY WORK IN NEW MEXICO
Although interest in cumulus clouds and thunderstorms was high
among the members of the Research Group in 1948, the cumulus season
passed in the vicinity of Schenectady without any significant flights
having been carried out. It was realized that the best results could be
obtained from the seeding of cumulus clouds in a region where storms
‘Cumulus Studies SANG} —
originate, rather than in a region which, like the Schenectady area, is
traversed by storms. Chairman Stine of the Operations Committee had
had experience as a forecaster in New Mexico, and he strongly recommended
that that region be used as a base for experiments with cumulus clouds. This
recommendation was seconded by Schaefer, who knew of the work being done
in this field by Dr. E. J. Workman’s group at the New Mexico School of Mines
and who had obtained a promise of co-operation from Workman.
Accordingly, it was decided to attempt a flight to Albuquerque, New
Mexico, to determine whether the radar and other facilities of Dr. Workman’s
group would be of assistance in this respect. In view of the waning cumulus
season even at that location, preparations were made to carry out full-scale
tests if proper clouds were formed.
As a result, members of the project spent three days at Albuquerque
during mid-October of 1948. A working arrangement was quickly made with
Dr. Workman and his staff for radar tracking and photography of the tests
to be made. Two seeding flights were made, one on October 12 and the
other on October 14. The second of these two flights was performed under
such satisfactory conditions that the results obtained were considered partic-
ularly significant.
For example, an exceptionally complete aerial photographic record was
made of the conditions of the cloud that was seeded from one of the planes,
including 176 photographs 4'' x 5", plus pictures taken every 45 seconds of a
group of instruments giving time, altitude, air speed, heading of the plane, and
other pertinent information. Every time a photograph was taken of the cloud,
another picture would be taken of a clock and other instruments, thus recording
when the photograph was taken amlother significant data. In this way an in-
valuable flight record was made of the test.
Further data were collected on the ground. Lapse-time movies were
made of the clouds as seen from the station, as well as a series of still
pictures, and radar was used to detect any rain that might fall. Although
some excellent supporting data were thus obtained, unfortunately it was not
as complete as it might be because of a failure of the radio communication
between the airplane and the radar station. But significant radar observa-
tions were made, and photographs were taken of the radar scope, giving a
complete set of records of radar observations for a considerable period
of time,
Four seeding operations were conducted on the October 14 flight. The
details of these seedings and the results obtained were discussed at consid-
erable length by Langmuir in an occasional report, (20) But a summary of
his findings is to the effect that rainfall was produced over an area of more
Cumulus Studies EAs
than 40,000 square miles as a result of the seeding--about a quarter of
the area of the State of New Mexico. And substantially all of the rain for
the whole of New Mexico that fell on October 14 and 15 was the result of
the seeding operations near Albuquerque on October 14. ‘“The odds in
favor of this conclusion as compared to the assumption that the rain was
due to natural causes are many millions to one.’’
An early estimate by Langmuir was that about 100,000,000 tons of
rainfall was produced. Later, using the rain reports from 3380 stations
given ina U.S. Weather Bureau publication, he concluded that the orig-
inal estimate was unduly conservative. 20) Said he: ‘“The evidence in-
dicated that the rain started from near the point of seeding shortly after
the time of seeding and then spread gradually at a rate which at no place
exceeded 22 miles per hour, over an area of at least 12,000 square miles
north to northeast of Albuquerque with an average of about 0.35 inches.
This corresponded to about 300,000,000 tons.’’
SILVER IODIDE AT NEW MEXICO
So satisfactory were the tests conducted at Albuquerque in 1948 that
it was decided to make a further study of cumulus clouds at that location
in the middle of July the following year. Much more elaborate plans were
made for this second expedition; for example, not one but a number of
airplanes took part, and virtually all the members of the Research and
Operations Groups went along.
Previous to the arrival of the main body of the project, Langmuir
and Schaefer investigated the general cloud situation in the various moun-
tain regions nearby and decided the cloud systems along the Rio Grande
Valley near Albuquerque were superior for their purpose to anything
they could find in other parts of Arizona and New Mexico. In addition,
the excellent radar, photographic, and shop facilities of the Experimental
Range of the New Mexico School of Mines appeared to be ideal for carrying
out the operations planned.
Between July 13 and July 22 a total of ten flights was conducted, on
eight of which two or three planes participated. Excellent co-operation
was enjoyed in every phase of the operation, and an extensive mass of
data was obtained both in the air and at the ground stations which were
set up. Seeding operations with varying amounts of dry ice and the
ground eee of a Silver-iodide generator were the subjects for the
flight studies (18
Again the dry-ice seeding was successful, and the results of the
various airborne seeding operations was quite satisfactory. But a new
factor was introduced into this second expedition which put an entirely
Cumulus Studies =43=
different aspect upon the results and had a tremendous influence on the
course of future investigations and analysis. This was the effect of ground
seeding with silver iodide.
As usual, close attention was paid to changes in weather conditions, in
order to obServe any correlation between such changes and the dry-ice seeding.
But, although Vonnegut was conducting some silver-iodide seeding on the
ground, this was disregarded by Langmuir, who was concentrating on the air-
borne dry-ice seedings. Consequently, when he noticed some weather conditions
which could not be explained by the airborne seeding, he was puzzled.
Then he suddenly became conscious of the fact that Vonnegut had been
trying to call the ground seeding of silver iodide to this attention, and he im-
mediately realized that this might explain the discrepancies he had observed.
Further study convinced him that this was, indeed, the case.
Not only that, but the results of the seeding activities in New Mexico the
preceding year were reconsidered in the light of this development. And it
appeared reasonable to conclude that the similar widespread effects produced »
in October, 1948, were the result of the silver-iodide seeding which was done
at that time, rather than of the dry-ice seeding, which had been the previous
interpretation.
Langmuir made, as was his habit, an exhaustive analysis of the available
data and presented a striking summary of his findings 18) from which the fol-
lowing is quoted:
“I wish particularly in this paper to describe the more wide-
spread effects that were produced by the operation of the silver-
iodide generator on the ground during July, 1949, near Albuquerque.
The first seeding with silver iodide during this stay in New
Mexico was on July 15, 1949, but the generator was not run for
more than a couple of hours on each day thereafter until the 19th,
when it was operated for a short time only, late in the afternoon.
On July 20 it was not operated at all, but on the 21st it was op-
erated for 13 hours, starting about 5:30 a.m. and using 300
grams, or a total of 2/3 pound of silver iodide.
‘*Tests made by Dr. Vonnegut have shown that each gram of
silver iodide dispersed under these conditions produced 107°
sublimation nuclei that are slowly effective at -5°C but very
rapidly effective at -10°C.
Cumulus. Studies -49 -
‘‘The new probability theory....has served a valuable guide
in devising an objective method of evaluating the distribution
in space and time of the rain which follows the operation of
the silver-iodide generator on the ground or in the airplane
flights near Albuquerque. To illustrate the results, we will an-
alyze the data obtained on two days, October 14, 1948 (Flight
45) and July 21, 1949 (Flight 110).
‘These days were chosen because large amounts of sil-
ver iodiue were used, but no seeding was done on the imme-
diately preceding days. Furthermore, the wind direction on
both days was rather similar. On both days the Weather
Bureau predicted no substantial amount of rain. Both morn-
ings were nearly cloudless, and on both days SW winds pre-
vailed from the cloud bases at 12,000 feet up to 20,000 feet.
At lower and higher altitudes and later in the day there were
also winds from the E, W, and NW. On both days, visual
effects indicating thunderstorms and heavy rain over wide
areas were observed a few hours after the start of the seeding
operations.
‘In the July operation our techniques had been improved
compared to those of the preceding October. In October ra-
dar observations covered only a period of about an hour in
the afternoon, for at that ime it was not suspected that the
rain that lasted well on to the morning of the 15th had any-
thing to do with the seeding.
*‘On July 21, 1949, however, we had complete radar cover-
age from early in the morning until late at night. Photographs
of the clouds were taken not only from planes but from the
ground, including lapse-time motion pictures with photographs
every few seconds. ;
‘Shortly before 8:30 a.m. on July 21, 1949, a single large
cumulus cloud began to form about 25 miles S of the field sta-
tion near Albuquerque in a sky that was otherwise cloudless.
This cloud was located near the Manzano Mountains, and the
silver-iodide smoke had been blowing from the N about 10 mph
so that it should have reached the position of the cloud.
‘‘Between 8:30 and 9:57 the cloud grew in height slowly at
the uniform rate of 160 feet per minute. At 9:57, when the top
of the cloud was at 26,000 feet (temperature -23°C), the upward
velocity of the top of the cloud increased quite suddenly, so that
the cloud rose 1200 feet per minute until at 10:12 it had reached
44,000 feet (temperature -65°C).
Cumulus Studies L'5(0)=
_ **At10:06, when the top of the clowl was 36,000 feet (temperature
-49°C), the first radar echo return was obtained from the cloud
at an altitude of 20,500 feet (temperature -9°C). The distance
given by radar was 25 miles at an azimuth of 165°, which was
exactly where the cloud was found to be from visual observations.
The area p precipitation in the cloud was about one square mile
at that time and was deep within the mass of the cloud. Within
four minutes, the precipitation area had increased to seven
Square miles, and within six minutes after the first echo ap-
peared, the precipitation had extended upward to 34,000 feet,
where the temperature was -43°C.
“The chain reaction in this cloud started at low altitude at
a time and place which agreed well with the trajectory of the
silver -iodide smoke.
‘*The first flash of lightning was seen at 10:10, four minutes
after the first radar echo was detected. In all, perhaps a dozen
flashes of lightning formed from this cloud, and very heavy rain
was seen to fall to the ground. Ihe top of the cloud moved to-
wards the W, but the lower part of the cloud, from which the
rain was falling, moved gradually to the NE.
*‘At 10:45, a second cloud about eight miles still further
to the NE developed a radar echo, and from that time on during
the day there was an increasing number of rainstorms giving
very heavy showers in the neighborhood. During the late after-
noon 1.2 inches of rain fell at the station where the generator
was located. The phenomena observed near and at Albuquerque
from the ground and the radio reports of exceptionally heavy
rain at Santa Fegave immediate evidence of the success of this
operation in producing heavy rain.’’
Langmuir’s report then analyzes river flow data and rain gauge data
for the region. In discussing the rain gauge data, he says:
‘The Weather Bureau observer with Project Cirrus in
New Mexico stated that he considered it possible or even prob-
able that seeding operations carried on there could have in-
creased the naturally occurring rain by five per cent, but certainly
not more than 10 per cent. If this were ture, it would be possible
to conclude that seeding operations have economic value only if
experiments are carried on many hundred of days, and a statis-
tical analysis is made of the rainfall data for all of these oper-
ations.
Cumulus Studies ails
‘*The rainfall data actually show, however, that the rainfall
on both October 14, 1948 and July 21, 1949 was exceptionally
high and could not have possibly been accounted for as the re-
sult of naturally occurring rain. This proof is made by the
analysis described in this paper.
‘‘The map of the State of New Mexico, which represents
about 120,000 square miles, was divided into eight octants or
45° sectors radiating out from Albuquerque. Then concentric
circles having radii of 30, 75, and 125 and 175 miles were
drawn on the map. This divided the whold state into 27 regions
whose average distances and directions from Albuquerque were
known.
‘‘By entering on the map for each of these regions the
average rainfall for Flights 45 and 110, a comparison could be
made of the distribution of the rain on those two days. An ob-
jective way of evaluating the similarity between such two dis-
tributions is to employ the statistical device known as the
correlation co-efficient. This was found in this case to be
+0.78+ 0.076. The chance that such a high value would occur
among these figures if one set of them were shuffled giving
a random distribution is only 1 in 10. Such close agreement
in the distribution on two days could thus hardly be the result
of chance. There must be an underlying cause.
‘“We believe that the close similarity in distribution is
dependent not only on the rather uniform synoptic situations
over the states that prevailed on these days, but also depended
on the fact that on both days the probability of rainfall depended
on the nuclei that spread radially out from Albuquerque, the
concentration decreasing as the distance from Albuquerque in-
creased.
‘*The next step was to investigate just what characteristics
of this distribution were so similar on these two days. On each
of the two days, nearly all of the rain that fell occurred within
four of the eight octants. If each sector were divided into four
to six parts arranged radially so that each would contain equal
numbers of observing stations (about eight per region), the an-
alysis showed that the average rainfall rose rapidly to a max-
imum in intensity about 30 miles from the point of seeding and
that in each of the four sectors it decreased regularly as the
distance from the source of the silver-iodide smoke increased.
Cumulus Studies 5h
‘In fact, this decrease followed quite accurately equations (2)
and (3), which indicated that the rain fall depended on the con-
centration of nuclei, and this, in turn, varied inversely in pro-
portion to the distance from the source.
‘‘This analysis makes it possible to separate the effects
of the artificial silver-iodide nuclei from that of the background
of sublimation nuclei that were already present in the atmosphere.
The analysis gave proof that C, = 0, so that there was no appreci-
able background on each of these two days. We must conclude
that nearly all of the rainfall that occurred on October 14, 1948
and July 21, 1949 was the result of seeding.
‘*The agreements between the intensity of the average rain-
fall in separate regions and the theoretical equations were So
good in each of the four sectors on October 14 and July 21 that
the probability factors for each sector ranged from 102 to 10%.
Taking all the octants together, the probability factor rose to ;
about 10 &to.l. 4
‘*For each of the eight octants that gave appreciable rain,
the rain started progressively later as the distance from the
source of the silver iodide increased. The advancing edge of
the rain area thus moved from Albuquerque on July 2l ata
velocity of about 15 mph and on October 25 at a speed of about
25 mph. These velocities agree well with the wind velocities
observed at various altitudes.
**The method of correlation coefficient can be applied to
the relation of the time of the start of the rain to the distance
from Albuquerque. This indicates that there is another prob-
ability factor which is the order of 108to l.
‘“Taking these results altogether, it seems to me we may
say that the results have proved conclusively that silver -iodide
seeding produced practically all of the rain in the State of New
Mexico on both of these days.
‘‘T have not mentioned what happened on the other days.
The results, although somewhat more complicated due to the
overlapping of the effect of seeding on successive days, are
almost as striking as those of Flights 45 and 110, in which we
used silver-iodide seeding. Very high probability factors are
found, which help confirm the results indicated by the analysis
of Flights 45 and 110.
Cumulus Studies = a=
‘*The total amounts of rain that fell in the state on the
two days as a result of seeding were found to be 800 million
tons on October 14, 1948 and 1600 million tons on July 21, 1949.
If these units are not so familiar to you, 1 may say that on
October 14, 1948, the total amount of rain resulting from seed-
ing was 160 billion gallons and on July 21, 1949, 320 billion
gallons.
‘Dr, Vonnegut has measured the number of effective sub-
limation nuclei produced by the type of silver-iodide smoke
generator used in our New Mexico experiments for each gram
of silver iodide used....One thus finds that, to get a 30-percent
chance of rain per day within a given area in New Mexico, the
cost of the silver iodide is only $1. for 4000 square miles.
‘If similar conditions prevailed over the whole United
States, the cost per day to double the rainfall would be only
of the order of a couple of hundred dollars. This verified an
estimate that I made in November, 1947 in an address before
the National Academy of Sciences that ‘a few pounds of silver
iodide would be enough to nucleate all the air of the United
States at one time, so that it would contain one particle per
cubic inch, which is far more than the number of ice nuclei
which occur normally under natural conditions.’ Such a dis-
tribution of silver-iodide nuclei ‘in the atmosphere might
perhaps have a profound effect upon the climate.’ ”’
The report then discusses a new theory which Langmuir had devel-
oped of the rate of growth of snow crystals in supercooled clouds contain-
ing known numbers of sublimation nuclei. After a brief exposition of the
basis of this theory, he says:
‘From the probability theory of the growth of showers
from artificial nucleation, one obtains the result that the
total amount of rain produced by operating a ground generator
increases in proportion to the square of the amount of silver
iodide used. Thus, with three times as much silver iodide
one would get nine times the rainfall. The intensities of the
showers would be no greater, but they would extend over a
greater area.
‘An analysis of the July 1949 rainfall in New Mexico,
Arizona, Colorado, Oklahoma, Kansas, and Texas gives evidence
that a band of heavy rain progressed in an easterly direction
during the period of July 20 to July 23 from southern Colorado
across the southern half of Kansas, where it gave 3 to 5 inches
Cumulus Studies ~54-
‘of rainfall in many places. It may have been dependent on the
silver-iodide nuclei generated near Albuquerque between July
18 and 21 and in central Arizona between July 19 and 21.
‘‘Furthermore, the heavy rains that spread throughout New
Mexico from July 9 to 13 before the start of Project Cirrus
seeding experiments appear to have depended on silver-iodide
seedings in Arizona on July 5 and 6.
‘Tt is very important that regular tests on certain selected
days of each week be carried out throughout the year, using
amounts of seeding agents just sufficient to obtain conclusive
statistical data as to their effectiveness in producing widespread
rain. It is to be expected that the results will vary greatly in
different parts of the country, because of the changes in synoptic
situations.’’
The significance of the two test projects at New Mexico is thus apparent.
They indicated not only the possibilities of silver-iodide seeding from the ground,
but they suggested a widespread effect on the weather of the nation. And, asa
result, the project conducted some experiments in periodic seeding which were
destined to have a profound--and controversial--significance.
VU - PERIODIC SEEDING
NEW MEXICO WORK
By this time, a rather close liaison had been established with Dr.
Workman and his co-workers at the New Mexico School of Mines. So,
in view of the significance of Langmuir’s analysis of the effects and
possibilities of silver-iodide ground seeding, and in order to test as
soon as possible his ideas on periodic seeding, a schedule of operations
on this basis was estiblished without further ado at New Mexico.
Starting in December, 1949, a silver-iodide ground-based gener-
ator was operated in New Mexico by the project on a schedule so planned
as to introduce, if possible, a seven-day periodicity into the weather
cycles of the nation. This schedule of regular weekly periodic seedings
used about 1000 grams of silver iodide per week, and it continued with
a few modifications until the middle of 1951.
Data were gathered by Falconer, and almost immediately Langmuir
found evidences of a definite weekly periodicity in rainfall in the Ohio
River Basin. Again, he conducted an exhaustive analysis of the facts and
performed elaborate mathematical calculations to determine the prob-
abilities that these variations in weather could have taken place by pure
chance.
He reported his findings and his conclusions to the National Academy
of Sciences, October 12, 1950 to the American Meteorological Society of
New York City on January 30, 1951 (25). and also to the New York Academy
of Sciences on October 23, 1951.(24) He pointed out that, during 1950,
there was a marked and statistically highly significant seven-day perio-
dicity in many weather elements. The significance was So high, said he,
that it could not be explained on the basis of chance; it could not have
occurred anyway from natural causes. It involved not only rainfall but
also pressure, humidities, cloudiness, and temperatures over much of
the United States.
In his paper to the New York Academy of Sciences, (24) Langmuir
said:
“‘Almost immediately, that is, during December 1949 and
January 1950, it was noted that the rainfall in the Ohio River
Basin began to show a definite weekly periodicity. A conven-
ient way of measuring the degree of periodicity was to calcu-
late the correlation coefficient CC between the rainfall on the
successive days during a 28-day period, with the sine or the
cosine of the time expressed as fractions of a week, the phase
being taken to be O on Sundays.
Periodic Seeding -06=
‘Just before the start of the periodic seedings, the corre-
lation coefficient CC(7) based on the seven average values for
the successive days of the week of the 28-day period amounted
to only 0.28, but in the next 28-day period the value of CC(7)
rese-to,0 017
*“Table I gives the average rainfall in inches per station
day during 140 days at 20 stations designed as Group A in the
Ohio Valley Basin, representative of an area of about 600,000
Square miles, The successive rows correspond to five succes-
Sive 28-day periods. It will be noted that the average rainfall
on Monday was 0.272", whereas on Saturday it was only 0.064",
a ratio of 4.3:1 The next to the last column gives CC(28), the
periodic correlation coefficients for each 28-day period, and
the last column gives the phases in the successive periods.
Taking the 35 separate values for the 4-week averages given
in the table, one gets CC(35) = 0.689 with a phase of 1.60 days.
This result is statistically highly significant.
*“These periodicities in rainfall were evident at almost
any set of stations in the northeastern part of the United States.
Table 2 gives the rainfall on successive Tuesdays and Saturdays
during a 12-week period during the winter of 1949-1950 at Buffalo,
Wilkes-Barre, and Philadelphia. This periodicity is almost the
Same as that found in the Ohio River Basin but with a one-day
phase lag. The striking contrast between the total rains on
Tuesdays and Saturdays runs parallel to the total number of
days on which rains of 0.1" or more occurred on Tuesdays and
on Saturdays.
‘‘Maps have been prepared giving for 24 successive 28-day
periods the distribution of correlation coefficients, CC(28),
among 17 subdivisions of the United States, these data being
based on daily weather reports of 24-hour rainfall at 160 sta-
tions. During the first five 28-day periods there were always
several adjacent subdivisions that showed high weekly perio-
dicities in rainfall. After May 1950, however, the periodicities
became somewhat sporadic, although highly significant perio-
dicities over large areas still occurred during more than half
of the periods after July 1950. Presumably the large amount of
commercial silver-iodide seeding in the western states (not done
with a weekly periodicity) masks the effects of the periodic
seedings in New Mexico. By a map, the areas were shown in
which known seeding operations have been carried on in 1951.
~
-102-
Appendix IV
Bibliography
Project Cirrus Reports
(75) First Quarterly Progress Report, July 15, 1947.
(76) Second Quarterly Progress Report, November 15, 1947.
(77) Third Quarterly Progress Report, February 15, 1948.
(78) Fourth Quarterly Progress Report, July 1, 1948.
(79) Fifth Quarterly Progress Report, September 15, 1948.
(80) Final Report, Contract W-36-039-SC -32427, December 31, 1948.
(81) Sixth Quarterly Progress Report, January 28, 1949.
(82) Seventh Quarterly Progress Report, March 15, 1949.
(83) Eighth Quarterly Progress Report, June 15, 1949.
(84) Ninth Quarterly Progress Report, September 15, 1949.
(85) Tenth Quarterly Progress Report, December 30, 1949.
(86) Eleventh Quarterly Progress Report, March 30, 1950.
(87) Twelfth Quarterly Progress Report, July 30, 1950.
(88) Thirteenth Quarterly Progress Report, October 30, 1950.
(89) Fourteenth Quarterly Progress Report, January 30, 1951.
(90) Fifteenth Quarterly Progress Report, April 30, 1951.
(91) Final Report, Contract W-36-039-SC-38141, July 30, 1951.
-103 -
Appendix IV
Bibliography -
Miscellaneous
(92) ‘‘Report of Cloud-seeding Experiments in the San Diego County and
the Santa Ana River Watershed’’; revised edition June 10, 1952, pub-
lished by John A. Battle, consulting meteorologist, Beaumont, Cali-
fornia.
(93) ‘‘Weather Control and Augmented Potable Water Supply’’; Extracts
from hearings before subcommittees of the committees on Interior
and Insular Affairs; Interstate & Foreign Commerce; and Agriculture
& Forestry; United States Senate, 82d Congress, First Sessions; on
SeOwoeacc, ands. oon Washington. DL ©. March i4s 15.16. 19. and
April 5, 1951; U.S. Government Printing Office.
DISTRIBUTION LIST
Additional copies of this report will be supplied
to qualified persons upon application to:
Research Publication Services
Room 2E38, The Knolls
Schenectady
GENERAL ELECTRIC ADVISORY COMMITTEE
R.J. Cordiner, New York H.A. Winne, Schenectady
P.D. Reed, New York N.M. DuChemin, Schenectady
H. V. Erben, Schenectady J.L. Busey, New York
R.W. Johnson, New York D.L. Millham, Schenectady
R, Paxton, New York L.R. Boulware, New York
J.W. Belanger, Schenectady C.H. Lang, New York
H.F. Smiddy, New York R.H. Luebbe, New York
C.G. Suits, Schenectady
RESEARCH GROUP
I, Langmuir, Schenectady
V.J. Schaefer, (12) Schenectady
B, Vonnegut
Arthur D. Little, Inc., Cambridge, Mass.
R.E. Falconer, Schenectady
K, Maynard, Schenectady
R.L. Neubauer, Schenectady
R. Smith-Johannsen, Waterford
7, ©). Blanchard
a Woods Hole Oceanographic Inst., Woods Hole, Mass.
M. Geller
Massachusetts Institute of Technology, Cambridge, Mass.
Charles Woodman, Meter and Instrument Dept., West Lynn
Arthur Parr, Schenectady
George Blair, Malta
STEERING COMMITTEE
Evans Signal Laboratory, Belmar, New Jersey
Michael J. Ference, Jr.
C.J. Brasefield
Committee on Geophysical Sciences, Office of Naval Research, Wash. D.C.
E.G, Droessler
Commander R.A. Chandler
Lieutenant Max A. Eaton
Commander G.D. Good
c/o Chief of Staff, U.S. Air Force, Washington, D.C.
Major P.J. Keating Colonel N.C. Spencer
Lieutenant Colonel J. Tucker
OPERATIONS COMMITTEE
Office of Naval Operations, Washington 25, D.C.
Commander Daniel F. Rex
Electronics Squadron, Griffiss Air Force Base, Rome, New York
Captain J.A. Plummer
Squier Signal Laboratory, Ft. Monmouth, New Jersey
OTHER
Samuel Stine
Colonel Rudolph C. Koerner, Jr.
Eugene Bollay
North American Weather Consultants, 1187 No. Green St.,
Pasadena, Calif.
Dr. Irving P. Krick
Water Resources Development Corp., 460S. Broadway, Denver, Colo.
William Hartnett
Weathercasts of America, 2123 Railway Exch. Building, St. Louis, Mo.
Professor Carl G. Rossby
Institute of Meteorology, 103 Flemingatan, University of Stockholm,
Stockholm, Sweden
Professor P.E. Sheppard
Dept. of Meteorology, Imperial College, Exhibition Road, London, nee
Professor R. Sanger
Polytechnic Institute, Zurich, Switzerland
Lieutenant (JG) W.E. Hubert
Institute of Meteorology, University of Stockholm, Stockholm, Sweden
. Hod Wells, AGC
Project Scud, Com. Air. Lant., Atlantic Fleet Weather Control,
Norfolk, Va.
Commander C.E. Tilden
Project Scud, Com. Air. Lant., Atlantic Fleet Weather Control,
Norfolk, Va.
Joseph Silverthorne
c/o Empressa Dean, Tegucigalpa, Honduras
Maurice Halstead
c/o Johns Hopkins Climatological Laboratory, Seabrook, New Jersey
Joseph B. Dodge
Gorham, New Hampshire
Alan Bemis
Massachusetts Institute of Technology, Cambridge, Mass.
Millikin & Farwell, Inc.
1002 Whitney Building, New Orleans, La.
Wallace E. Howell Associates
Cambridge, Mass.
Jack Barrows
Chief, Fire Research Region I, U.S. Forest Service, Missould, Mont.
Dr, Marcel de Quervain
Institute for Study of Snow and Avalanches
Davosdorf/Weissfluhjoch, Switzerland
OTHER (continued)
Signal Corps Engineering Laboratories, Fort Monmouth, N.J.
General E.R. Petzing
Dre A Zaki
Dr. M. Ference
S.E. Petrillo
E.L. Nelson
Bd. Fister
Office of the Chief Signal Officer, Washington 25, D.C.
Major General G. 1. Back
Brigadier General J. D. O’Connell
Colonel C.J. King
G-E LIBRARIES
Nela Park Library, Cleveland
Electronics Park Library, Syracuse
Main Library, Schenectady
William Stanley Library, Pittsfield
Thomson Laboratory Library, River Works
F.E. Arnold, West Lynn Library
Chemical Division Library, Pittsfield
Aircraft Gas Turbine Library, Lockland, Ohio
Household Refrig. Library, Erie
Switchgear Department Library, Philadelphia
Loco. and Car Equip. Dept. Library, Erie
Fort Wayne Library
KAPL Library
Johnson City Library
Project Hermes Library, Campbell Ave. Plant, Schenectady
Research Laboratory Library, The Knolls
Accessory Turbine Library, 63NG, River Works
G-E PERSONNEL
ELECTRONICS DIVISION
Syracuse
W.R.G. Baker V.M. Lucas
L.R. Cohen G.F. Metcalf
Coe DeVviore J.W. Nelson, Jr.
J.J, Farrell lain, Ollichnlellels die.
ae, Melanie iC, Rives
H.F. Konig
GENERAL ENGINEERING LABORATORY
Schenectady
IAG Rea
NEWS BUREAU
Schenectady
James Stokley
News Bureau
PUBLIC RELATIONS SERVICES DIVISION
Schenectady
B.S. Havens (5)
N.B. Reynolds
ENGINEERING SERVICES DIVISION
Tech. and Eng. Adm. Services Dept., Schenectady
J. Horn
General Technical Services Section, Tech. Data Center, Schenectady
BER. Vv rooman ct aan
General Engineering Laboratory, Schenectady
ie Ae RIC Hh ET ET
Process Tech, Program, Technical Education, Room 194, Building 6,
Pittsfield
LEGAL AND PATENT SERVICES DIVISION
Patent Services Department, Schenectady
PA. Frank
RESEARCH SERVICES DIVISION
Research Laboratory, The Knolls, Schenectady
R.W. Bengtson
D.E, Chambers
R, Damon
J.R, Eshbach
M.D. Fiske
V.H. Fraenckel
M.H. Hebb
J.H. Hollomon
RW; Larson
H.N. Leifer
Mood “Ozerott
EM. Pell
R.W. Redington
W.A. Thornton
W.W. ‘Tyler
V.C. Wilson
J.R. Young
NOs Or:
Copies
00
50
30
Transportation Officer, SCE L, Evans Signal Laboratory, Bldg. 42,
Belmar, New Jersey. Marked: ‘‘For Signal Property Officer’’.
Chief of Naval Research, Navy Department, Washington 25, D.C.
Attn: Code N428
Chief of Staff, U.S. Air Force, Washington 25, D.C.
Attn: Director of R&D, DCS/M
Director, Mt. Washington Observatory
2 Divinity Avenue, Cambridge 38, Massachusetts
Committee on Geophysical Sciences, Research and Development
Board, Washington 25, D.C.
Panel on Meteorology, Research and Development Board
Washington 25, D.C.
U.S. Weather Bureau, 24th and M Streets, N.W., Washington 25, D.C,
Attn: Dr. H. Wexler
U.S. Weather Bureau, 24th and M Streets, N.W., Washington 25, D.C.
tins Digsnoss Gunn
U.S. Weather Bureau, 24th and M Streets, N.W., Washington 25, D.C.
Library
Director, National Bureau of Standards, Washington 25, D.C.
Attn: Mr. Hugh Odishaw
Department of Agriculture, Washington 25, D.C.
Commissioner, Bureau of Reclamation, Washington 25, D.C.
Attn: Section 724
Chief Hydraulic Engineer, U.S. Geological Survey, Washington 25, D.C.
Attn: Division of Water Utilization
Director of Aeronautical Research, N.A.C.A., 1724 F Street, N.W.,
Washington, D.C.
No. of
Copies
N.A.C.A. Laboratories, Cleveland Airport, Cleveland, Ohio
Attny Mx, L.A, Rodert
1 Massachusetts Institute of Technology, Department 6f Meteorology,
Cambridge 39, Massachusetts, Attn: Dr. H.G. Houghton
1 New York University, Department of Meteorology, New York, New York
Attn: Dr. B. Haur witz
1 University of Minnesota, Institute of Technology, Minneapolis 14, Minn.
Attn: Athelstan F. Spilhaus, Dean
1 University of Chicago, Department of Meteorology, Chicago, Illinois
Attn: Dr, Horace Byers
il University of California at Los Angeles, Department of Meteorology,
Los Angeles, California, Attn: Dr. M. Neiburger
iL Pennsylvania State College, Division of Meteorology, State College,
Pennsylvania, Attn: Mr. H. Neuberger
if Director, Blue Hill Observatory, Milton 86, Massachusetts
i Institute for Advanced Study, Princeton, New Jersey
Attn: Dr. J. Von Neumann
4
}
cl Woods Hole Oceanographic Institution, Woods Hole, Massachusetts
Ating sor. @7). Iselin
1 St. Louis University, 3621 Olive Street, St. Louis 8, Missouri
Attn: Dr. J. B. Macelwane, S.J.
it Electrical Engineering Res. Lab., The University of Texas, Box F,
University Station, Austin 12, Texas
1 New Mexico School of Mines, Box 6000, Station A Seas ae
New Mexico, Attn: Dr. E.J. Workman
1 University of New Mexico, Albuquerque, New Mexico, Attn: Dr. V.H. Regen
i Scripps Institute of Oceanography, La Jolla, California, Attn: D. Leipper
il Cornell Aeronautical Laboratory, P.O. Box 235, Buffalo 21, New York
Attn: Dr: S. Chapman
No. of
Copies
iL
3
Chief, State Water Survey Division, Urbana, Illinois
Chairman, Operations Committee, PROJECT CIRRUS, General
Electric Hangar, Schenectady County Airport, Schenectady
Mr. Louis W. Jolliff, 1108 West Columbia, Champaign, Illinois
Professor Thompson, Meteorological Head Office, 315 Bloor Street
West Toronto, Ontario
Mr, E.L. Davies, Deputy Chairman, Defense Research Board,
Ottawa, Ontario
Mr. G.W.C. Tait, Suffield Experimental Station, Suffield, Alberta
New York University, College of Engineering, University Heights,
New York 53, New York, Attn: Dr. H.K. Work, Director of Research
U.S. Department of Agriculture, Division of Fire Research,
Washington 25, D.C., Attn: A;A. Brown, Chief
University of Alaska, College, Alaska, Attn: Dr. E.F. George
Commercial Service Section, Syracuse, New York, Attn: C.P. Reynolds
Armed Forces College, Norfolk 11, Virginia
Chemical Corps Technical Command, Army Chemical Center,
Maryland, Attn: Dr. Solomon Love
Mr. H.F. Huddleston, U.S. Department of Agriculture, Bureau of
Agriculture Economics, Division of Agriculture Estimates,
Washington, D.C.
American Meteorological Society, Boston 8, Massachusetts
Bureau of Aeronautics, Project AROWA, Norfolk, Virginia
PR@ ARATE,
Mp 4s
Fy
Be
J-5BA
GENERAL@Q ELECTRIC =
Kescarch Laboe
SCHENECTADY, N. Y. |