Proceedings o f The South African Sugar Technologists' Association-June 1972
A REVIEW OF RAINFALL STIMULATION BY
MEANS OF CLOUD SEEDING AND ITS POSSIBLE
APPLICATION IN THE NATAL CANE BELT
By E. HUGHES
"Ellesmere'?,P. 0.Knapdaar, Cape Province.
Introduction supercooled cloudy air, and silver iodide
The modern history of rainfall stimulation by generators provide about 10'3 nulcei per gram of
means of cloud seeding began in 1945 with the smoke at -IOOC, only a few grams of silver iodide
discovery that various substances, notably dry ice, would be required per cloud to provide one ice
could cause super-cooled vapour clouds to con- particle per litre. One litre of cloudy air was
dense into large droplets. In the following year it assumed to have about one million cloud droplets
was found that silver iodide crystals had the same (10 - 20u diameter) and these should condense to
effect on super-cooled clouds, causing them to pre- form one rain drop about Imm in diameter. It
cipitate ('3). Super-cooled clouds are those which was assumed that this would upset the colloidal
contain water vapour and liquid water at stability of the cloud. The growth of cloud
temperatures below OOC, and these occur in all droplets to precipitation size is explained by
parts of the world. Great enthusiasm greeted these reference to the Bergeron Findeison process where
discoveries and wild claims were made regarding the ice particles will grow at the expense of water
man's ability to control the weather and to end droplets due to the lower vapour pressure over
droughts. However, several large scale projects i~e(~,13). No consideration was given to the
yielded conflicting results with apparently large concurrent dynamic effects due to the release of
.increases in precipitation in some cases and latent heat of fusion or condensation, nor to the
decreases in others (",'3). effects upon the field of motion of the cloud.
In particular, Project Whitetop, conducted by The results from operations based upon the
Roscoe Braham (1966), and a 5 year Arizona static theory were often disappointing regarding
experiment, conducted by Louis Batten (1966), cloud growth and the yield of water (I3). The
yielded inconclusive results. In the case of Project dynamic theory involves the releasing of the
Whitetop, a 20% reduction in precipitation over maximum amount of latent heat of fusion and
an area of 100,000 square miles was recorded (13). condensation as rapidly as possible, thereby
Public and scientific interest in weather materially increasing in succession the temperature
modification waned throughout the world, partly of the cloud (by 0,5 to IOC), the buoyancy of the
as a result of the above experiments and also due cloud and finally the vertical and horizontal
to the failures of "get-rich-quick" operators, and dimensions of the cloud. This increases the life-
the whole subject fell into disrepute. time of the cloud and the precipitation from it.
Interest was revived largely by the successes More moist air is drawn up into the cloud and
of commercial operators who had persisted in the this in turn is condensed and eventually pre-
field in the U.S.A. and who were supplying extra cipitated as snow, hail or rain. The technique
water to large hydro-electric and irrigation dams. requires the release of about 20 pyrotechnic flares
Their well-documented operations prompted the yielding 50 grams of silver . iodide each, for the
authorities to press. for further research into the treatment of one cloud. This dynamic approach,
matter. In 1961-62, the American Congress voted which was largely developed by Prof. Joanne
$100,000 for research into "increasing rainfall by Simpson in Florida, has yielded positive and
cloud seeding" and this -r-dsearch is now largely co- convincing seeding results (13,14,'5,16).
ordinated by the U.S. Bureau of Reclamation, as The micro-physical : .dynamic theory is the
Project Skywater('), which is only one of several most recent and, in terms of extra water
weather modification projects with vast financial produced, is the most successful explantion of the
backing. This effort has led to a rapid increase in precipitation stimulation mechanism. This theory
knowledge in many fields of weather modification combines the most recent discoveries in the field
and the confusion and apparent contradictions of of cloud micro-physics with that of cloud
previous experiments have now to a large extent dynamics, and this approach appears to hold
been elucidated (1,2,3,). great promise and may soon lead to an optimising
Some theories of super-cooled cloud seeding of rainfall stimulation techniques (4).
In the original or static theory, it was
assumed that cloud microphysics should be altered The main results of Simpson's work
by the introduction of freezing nuclei into a Two important results emerge from Simpson's
supercooled cloud. Since an average cumulus work and her early use of numerical models.
cloud might 'contain about 10'3 litres of (a) Seedability vs seeding effect
Proceedings of The South African Sugar Technologists'Association-June 1972 197
Seedability is defined as "the predicted seeded climate, but most impressive work has also been
maximum cloud top height minus the predicted done by Robert Elliott of North American
unseeded maximum cloud top height in Weather Consultants, California, U S A . on
kilometers". orographic clouds. The records compiled by his
Seeding effect is defined as "the observed (by company over 20 years of supplying large
aircraft, airborne camera, radar etc.) maximum amounts of water to hydro-electric and irrigation
cloud top height minus the predicted unseeded dams in the Sierras have proved invaluable in
maximum cloud top height in kilometers". Results documenting the reawakened interest in weather
obtained in 1965 are shown in Figure 1. If the modification in the early 1960's ('I,'*).
model and data were perfect, all seeded clouds
(circles) would lie along the straight line with
slope one, since seedability and seeding effect
should be equal for seeded clouds. The unseeded
clouds (squares) should lie along the straight line
with slope zero, since no matter how high their
seedability the control clouds should not grow
above their predicted unseeded tops. In Fig. 1,
both the means and regressions of the seeded and
unseeded populations separate significantly into
two distinct populations. Seeding effect vs seed-
ability correlated ot 0,973 (p-0,01) for seeded
clouds and was about zero for unseeded clouds.
This demonstrates the precision of the model in
predicting seeded and unseeded cloud top heights 1 3 .
(b) The second Important result to emerge from
this work has been the recognition of four basic
regimes which will lead to different behaviour in
seeded clouds. These regimes are:
(i) Explosive growth mode: Seeding will lead to
considerable cloud growth and increased
(ii) Cut o f f tower mode: Seeding will lead to
(iii) N o growth mode: Seeding will lead to no
Increase or a decrease in precipitation.
(iv) Disturbed. mode: Seeding will probably
lead to no increase or a decrease in precipitation
on a naturally rainy day.
The classification of the day on the above Fig I Seedability versus seeding effect for 1965
basis is made on the results of the 0800 hrs. Caribbean cumulus seeding experiment (See text
radiosonde observation and a further observation for explanation
at 1200 hrs. Should a day be a seed day (mode i)
then a computation of the increased precipitation Some theories on warm cloud seeding
is made after measurement of cloud base, height Work on warm clouds, i.e. clouds warmer
and diameter (l7,14). than OOC, has not progressed as rapidly as that on
his work forms the basis for a sophisticated super-cooled clouds. An intensive effort is now
climatological survey method which has been being made to understand the processes involved
developed by Profs. Simpson and Garstang which, in the modification of these clouds and numerical
on the basis of analysis of meteorological records models are being developed which are already
and some field study, gives a reasonably accurate yielding reliable results. The seeding materials
picture of the weather modification potential for used for warm clouds are strongly hygroscopic
an area. This is a considerable advance and saving such as urea or finely ground salt (NaCI). The
in cost, showing the probable amount of water to insertion of the hygroscopic seeding material
be produced by seeding, the unit cost of the water causes the formation of a heirarchy of droplet
and the changes in streamflow due to the extra sizes and the mechanism of droplet growth is
water produced (7). thought to be primarily due to coalescence caused
A stratification of previous experiments (such by the different rates of fall of different sized
as Project Whitetop) has usually shown that the drop~.(~,'O).
decreases which were so baffling were due to Some consideration has already been given to
unfavourable synoptic conditions. In general, the a combination of the "warm cloud/cold cloud"
greatest response to seeding is on "fair" days with techniques where, in certain cases, a warm cloud
isolated showers (11,14). might be induced to grow into a super-cooled
Prof. Simpson's work has been done mostly cloud and then be treated accordingly ( 9 ) . When
with supercooled cumulus clouds in a maritime the techniques for treating warm clouds have been
198 Proceedings o f The South African Sugar Technologists' Association-June 1972
developed to the present stage of reliability of TABLE I
those used for treating then a Comparative costs of water produced from natural and
great deal more moisture will be available for artificially induced sources
The quantity of water produced by cloud seeding Water Supply Costs Source
and some cost/benefits to be derived from it Rand Costs (5)
The amount of water produced by weather per 1234 per m3
modification will clearly vary from area to area m3 cents
and from year to year. Some typical in-creases
quoted are as follows: Grahamstown Municipality 136.75 11
Wemmershoek Dam 40.85 3.3
Voelvlei Dam 70.70 5.7
(i) 137 million cubic meters of "new" water per Sewage Purification High 81.70 6.6
annum in the Upper Colorado River Basin Est. Low Est. 54.45 4.4
a t an estimated cost of R1,OO - R1,50 per Orange River Development
1234 m3, and a n estimated benefit of R20 - Project 7.075 0.574
R25 per 1234 m3. ( I , Vol. 1 p. 47). Weather modification High
(ii) an increase of 171,2 million cubic metres per Est. 2.15 0.176
annum in the reservoir to be obtained from a Water Low Est. 1.075 0.088 (1,2,3)
area of 25 888 sq km. (2, p. 316)
(iii) a 15% increase in stream flow "reasonably
anticipated" in the Colorado River Basin o r
11,000 sq miles a t a cost of R2,30 per 1234 Some costs and benefits are quoted in Table 2.
m3. (2, p. 433)
(iv) 407 220 m3 of extra water could be produced TABLE I1
per cloud in Miami during one normal
month (May) and this would produce Costs and Benefits of water derived from various sources
165 738 540 m3 over the area, o r a n increase
of 375mm. and a 23% total increase (13). Location Cost per cubic Cost benefit
(v) The anticipated increase for the Colorado metre ratio
River Basin Pilot Project is now estimated I. Colorado River 2295 million 1:lO (9)
a t 30%. (3, p. 95). Basin cubic m.
R1.11234 m 3
2. Australia 25,000 sq.km Cost per annum
In general it would appear that a t this stage, (Victoria) "treated" per acre less than
increases are of the order of 15% - 40% of annual I penny. (6)
rainfall (4). 3. South African 25,000 sq.km Cost S.A. 1,5c
Cloud mergers equivalent of 2 above "treated" per hectare.
'It may be noted here that most of the 4. The costibenefit ratio can be as high as 150. (9)
experimental work on cumulus clouds has been
done on isolated clouds. Simpson, however, has
considerable documentation on "cloud mergers", No critical assessment has yet been made of
where a seeded cloud has merged with a n the value of the extra water produced by the
unseeded one and the result has usually been a "extra-area effect" (discussed in the next section),
spectacular increase in water production. In one but it clearly is large and must be considered
case a merger produced 8797,6 acre ft of water, from the start of operations.
and it would have taken 36 isolated clouds t o For standing crops and general pastoral
produce the same amount ,of water. farming, the calculation of the value of the water
Work is proceeding actively along the lines of becomes even more complex. A point which has
inducing mergers by seeding numbers of clouds in been appreciated by the Australians is that the
rapid succession and so attempting t o induce the timing of the rainfall is often of great importance.
formation of squall lines, tropical storm bands It has been calculated that in the Malleen
and giant cumulonimbus systems, which are Wimmera wheat growing area of Victoria,
productive of enormous quantities of water. It Australia, an increase of 25 mm from August 1st
should be noted, however, that at this date no to October 31st would give a gain of A & I
sure technique has been developed to predict million at a cost of A & 6 000. (6). Also a n
whether a given day is any more favourable for increase which might be statistically non-
"mergers" than any other seedable day, or the significant might yield high financial returns. The
manner in which mergers can be initiated with rule of thumb .calculation for cane growing is that
certainty. 100 mm of water gives a n extra 9 tons of cane
Some comparative costs of water from per hectare. Consideration of the S-shaped growth
conventional sources a n d f r o m w e a t h e r curve common to most agricultural crops indicates
modification are given in Table 1. The that relatively small increases in precipitation
costlbenefits of seeding will vary greatly with the around about the mid-point, will give large growth
value and end use of the water produced. increases.
Proceedings of The South African Sugar Technologists'Associatio
Where hailstorms are frequent and damaging, 3.Persistence effects
the prevention of the formation of hail may be Briefly, the persistence effect is the tendency
economically as important as, or more important ,
for "rain to breed rain and drought to breed
than the production of more water. Severe hail drought". This seems to apply to large areas for
forming clouds can now be successfully treated by several seasons as well as the tendency for storms
cloud seeding (11,12). to follow the same fairly narrow path during a
season, and this phenomenon may be more
Some phenomena assosciated with cloud seeding apparent in dry rather than naturally moist
and their effect upon the evaluation of results and climates. This natural phenomenon was noted in
experimental work South Africa ( I * ) and Australia (6, 7) but to an
A number of phenomena are generally apparently lesser extent in America. The per-
associated with cloud seeding attempts and where sistence effect could prove very valuable in the
those have not been fully appreciated, or ignored, field of weather modification, but any experiments
they have seriously confused field studies and which do not take due note of this may well give
experimental results. In certain cases the conflicting results, as happened with early experi-
evaluation methods also were faulty or inadequate. ments with a randomized cross-over design for an
experiment in the Snowy Mountains in Australia
1. Evaluation techniques and methods
The earliest method of evaluating the success P).
or otherwise of seeding operations was by means
of a comparison of two supposedly similar climato-
logical areas, the clouds in the one area being Conclusions and recommendations
seeded and the clouds (and rainfall) in the other The benefits of obtaining greater reliability in
area being left as control. This method is what has hitherto been the greatest a n d . most
obviously fraught with many difficulties and important variable in agricultural production, the
uncertainties and many years of seeding would be rainfall, as well as a reasonable increase in preci-
required before any clear result emerged. pitation, are so obvious to the agricultural
Comparisons of stream flow records were used in community that they need not be belaboured here.
the early stages and these are still most valuable It is clear that cloud seeding can, under certain
in the evaluation of seeding results (4,17). conditions, and if properly conducted, produce
significant amounts of water very economically. In
order to take advantage .of this new knowledge in
The modern methods rely heavily on the use the most beneficial manner, the following lines of
of radar, which is calibrated by means of ground approach could be the most productive.
observations, to measure rainfall intensity and
duration ( 3 4, 5 , 6). The development of A potential user of water (farming
numerical computer models and their use for community, municipality, etc) could, after a
predictions of cloud growth and precipitation, thorough climatological survey of the area had
verified by observations on suitably randomised been carried out, consult with weather
clouds, have yielded a considerable amount of modification meteorologists and decide upon the
knowledge. optimum managed climate possible in their area.
In all probability this would not differ markedly
2.Extra-area ejrect or large scale dynamic ejj;fcts from a "good average year". Having jointly
A considerable problem for a long time has decided upon this optimum climate and having
been the question of whether extra water was in stated any special considerations regarding floods
fact being produced by seeding or whether water or the timing of rainfall which might apply, then
was merely being re-distributed. the cost of attaining this optimum for a period of
say 5 years could be calculated.
While the difficulties in the analysis of this
problem are obviously formidable, it became Having reached this point, the whole
apparent many years ago that, while large and pre- operation could be put on an "insurance" basis
dictable increases in precipitation were found in with, in this case, cane growers being com-
the target area, considerable increases in pre- pensated should the climate for any one year
cipitation were usually found up to 100 miles and deviate too markedly from the optimum.
more downwind of the target area, and that these
increases were as much as 10 times the amount In view of the shortage of water in South
that would normally have fallen there (=, 3 , 4, ',). Africa, and the greater and greater demands being
While the location and magnitude of these made on existing supplies, it is clear that many
increases or "extra area effects" cannot at this persons and public bodies will have an interest in
stage be predicted with as high a degree of weather modification and will want their share of
accuracy as the precipitation in the target area, the extra water being produced. This makes it
yet it would appear that the quantity of water imperative that potential water users, and in
produced by the extra area effect is considerably particular the farming community, should
greater than that produced for the target area. carefully examine this new technology and decide
This offers further support for the correctness of how best to use it, and how best to incorporate it
the dynamic theory of cloud invigoration (4). into their farming systems.
200 Proceedings of The South African Sugar Techno!ogistsl Association-June 1972
I.Anon., Atmospheric Water Resources Program "Project Proceedings of the First National Conference of Weather
Skywater". 1968 Annual Report. Vol. I and 11. United Modification sponsored by: Amer. Met. Soc. Albany, New
States Department of the Interior, Bureau of Reclamation, York, U.S.A. April - May 1968.
Division of Atmospheric Water Resource Management, Proceedings of the Second National Conference on Wea-
Denver, Colorado U.S.A. ther Modification sponsored by: Amer. Met. Soc. Santa
2. Anon., Atmospheric Water Resources Program, "Project Barbara, California, U.S.A. 1970.
Skywater' 1969 Annual Report. (address as in I. above). Simpson, Joanne., Cumulus Cloud Modification Progress
3. Anon., Atmospheric Water Resources Program, "Project and Prospects. Repr. "A Century of Weather Progress.
Skywater" 1970 Annual Report. )address as in I. above). Amer. Met. Soc. U.S.A. 1970.
4. Anon., Atmospheric Water Resources Program, "Project Simpson, Joanne., Seeding Cumulus in Florida: "New
skywater" 1971 Annual Report. (address as in I . above). 1970 Results" from Science.
5. Anon., Construction in Southern Africa. June 1970 Vol. Simpson, Joanne., On Cumulus Entrainment and One-
15. No. 3 p. 79. Dimensional Models. Repr. Journill of Atnlospheric
6. Bowen E.G., Cloud seeding Australian Science Journal Sciences, Vol. 28, No. 3. April 1971. pp 449 - 455.
1967, C.S.I.R.O., Sydney, Australia. Simpson, Joanne., How to Build Clouds and Bring Rain.
7. Bowen E.G., Review of Current Australian Cloud Seeding Farmers Weekly, Nov. 12 1971.
Activities C.S.I.R.O. Sydney Australia. Sopper, William E., Heimstra, L.A.V., Effects of Simu-
8. Garstang M., Personal Communication Brooks Museum, lated Cloud Seeding on Streamflow of Selected Watershcds
University of Virginia, Va. U.S.A. in Pennsylvania. Repr. Water Resources Bull., Journal of
9. Heimstra L.A.V., Kunsmatige neerslagvermeerdering: 'n the American Water Resources Assn. Urbana, Illinois
% sent per 4546 liters ekstra water? Die Siviele lgenieur 61801, U.S.A.
in Suid Afrika, Julie 1970. Whitmore, J.S., Technical Bulletin Report No. 25, 1962.
10 Heimstra L.A.V., Kunsmatige Vermeerdering van Reenval. Dept. of Water Affairs, Patterson Building, Pretoria,
Departement van Siviele Ingenieurswese Universiteit van Rep. of South Africa.
Stellenbosch. Februarie, 1970.