Prof.T.Shivaji Rao,

Director, Centre for Environmental Studies,

Gitam University, Visakhapatnam-530 045 (best website)
 [maps scale:1cm:2.5 km.] Excellent maps  [African Expts.]]  [Farmers conduct cloud seeding operatins to reap the maximum benefits to promote Self-help in local regions ]  (Thailand Expt)

(Full paper in original is enclosed as  second proof for the success of warm cloud seeding with common salt) 

Success will not only depend on the amount, application and type of seeding material used but also the prevailing weather and atmospheric conditions including cloud type, cloud temperature and water content. For example the 1979 to 1980 Victorian experiment was abandoned due to a lack of suitable clouds.

Determining success will also involve the establishment of a rigorous and appropriate experimental design. This is needed so as to be able to detect with reasonable certainty what changes in rainfall occur. Observations and measurements will be made by three nearby radars operated by the Bureau of Meteorology, surface rain gauges and by instruments attached to the research aircraft.

All air contains moisture. When the warm air rises from the surface of the earth due to the heat from the Sun the air rising into sky begins to cool and some of the moisture condenses on dust particles into small droplets that cause cloud formation. Clouds generally form within about 1 or 2 kms from the ground level depending upon the ambient temperature, water vapour, climatic conditions and topographical features.

It will be interesting to know that more than 99% of the cloud content is the air. Hence clouds contain only an average of one percent of the total atmospheric moisture at any time. So even if cloud seeding doubled the efficiency of a cloud to give rain, the cloud system would perhaps remove only about 2% of the available atmospheric moisture leaving the remaining 98% for other purposes. The amount of water vapour required for saturation of air at different temperatures is presented in the following table-1:





























It is well known that clouds begin to form in the sky when the air becomes super saturated and thereby forces the water vapour to condense on particles known as cloud condensation nuclei, (CCN) that grow into cloud drops. Ice crystals also form in those regions of the cloud where temperatures are below freezing levels while liquid droplets also co-exist in such subfreezing environment as Supercooled Liquid Water (SLW) that is the key for cloud seeding. But there is also warm cloud seeding process. Several scientists investigated on the various factors that influenced the formation of clouds and their precipitation. It was found that some clouds give rain when favourable conditions exist.

The water droplets in the clouds are so small (10 micro-meters in radius) that they cannot reach the earth without getting evaporated before touching the ground. Hence if large rain drops (500 to 2000 micro meters in radius) that can reach the earth without evaporating are to be formed, about a million small cloud droplets must join together and for this purpose mere adhesion and condensation are not enough. There are 2 theories about rain-drop formation, each of which is valid under different climatic conditions.

Rain drop formation and theories :

The first theory of rain-drop formation is based on “Langmuir chain-reaction” that mostly occurs in the hot and humid atmosphere of the tropics. The larger droplets fall much faster than the smaller ones such that bigger drops in a cloud overtake and absorb all the smaller drops found in their downward path. In a rising cloud, the up-currents hold the big drops in suspension while the small ones are caught up and merge with the big ones. If the large drops become too big, they burst into fragments that again collide and coalesce with small droplets and ultimately fall to earth as rain-drops.

The second theory of rain-drop formation is based on “Bergeron chain-reaction” that mostly occurs in the temperate zones. The clouds that precipitate consist at the higher levels, of ice crystals and supercooled water vapour. For the formation of big droplets in the cloud an ice phase is a precondition and for this transition from the liquid to solid phase a freezing nucleus is essential. The freezing nuclei allow supercooling of water droplets upto -15oC before ice formation occurs. Infact pure water suspended in air does not freeze until it touches about – 40oC temperature. The ice-crystals grow rapidly by absorbing all the super cooled droplets on their way down and large lumps of ice are formed and they melt at about zero degree centigrade level in the sky to become rain-drops that fall on earth. (Fig-1)

In warm clouds to be seeded there is a serious shortage of large water droplets over 20 microns in diameter and this reduces the efficiency of the cloud to give rain. In order to induce such clouds to give rain, the coalescence mechanisms have to be stimulated by injecting the clouds with the hygroscopic particles like sodium or calcium chlorides from an aircraft. The Langmuir chain-reaction gets a boost and the clouds which were formerly unproductive become productive and the clouds which would have given normal rainfall would give some additional rainfall. Similarly the dearth of ice-crystals in cold clouds is removed by seeding them with silver iodide to ensure one nucleus per litre of cloud air so that such seeding generates additional ice-crystals and thereby accelerates the rain-forming processes for providing additional rainfall.

Cloud seeding for artificial rains :

Since salt crystals are abundant in oceanic regions due to wave action and evaporation they favour larger cloud droplets that collide and coalesce to initiate rain fall well within the life time of the clouds. Oceans cover about 70% of the Earth’s surface and receive about 80% of the global rainfall. But the atmosphere over continental regions generally contains much smaller and more numerous condensation nuclei without much hygroscopic content like in the marine atmosphere and hence the medium sized clouds formed in such regions usually dissipate before coalescence mechanism has had a chance to initiate rain.

The augmentation of rain-fall from the warm clouds is based on the assumption that there is a deficiency of giant hygroscopic nuclei (GCN) to change cloud-drops into the raindrops and if salt particles are injected into suitable clouds, they will grow by condensation and then by collision and coalescence mechanisms that initiate rain-fall within the life time of the clouds.

Seeding agents :

Salt particles (NaCl, NH4NO3, urea) of about 5 to 20 microns if injected into the clouds grow rapidly due to their hygroscopicity into 40 to 100 microns very quickly and subsequently when these larger grown particles collide and coalesce with smaller cloud droplets precipitation develops. However salt seeding lacks the multiplying effect of silver iodide seeding. Salt particles are million times larger in mass than AgI particles. Salt particles are 100 times lesser in concentration. But in 1948, Langmuir advanced a chain reaction process that could multiply the seeding effect enormously. A rain drop of 5 mm to 10 mm diameter becomes hydrodynamically unstable and breaks into about 10 droplets. Since the terminal velocity of a large raindrop at break-up point is about 8 meters/second, it gets suspended at about a fixed level in an 8 meter/second updraft. After breakup the little drops move upward towards lower updraft velocities, growing by coalescence while moving up with the updraft until they become large enough to fall back. They stop again at the 8m/second level in the updraft and they breakup again and repeat the cycle. If the salt seeding increases the growth of a few large drops, then under these circumstances the precipitation can be enhanced and the extra rain thus generated reaches the ground when the updraft dies away in about 20 minutes.

Sometimes some cumulus clouds produce precipitation by a mixed phase inclusive of coalescence and ice processes. Hail storms are produced as hail-embryos by the coalescence process and then frozen. Premature seeding might start precipitation but it may reverse or restrict the natural growth of a cloud later in the day and hence premature seeding might advance the time of precipitation without necessarily increasing the total precipitation.

Hygroscopic seeding methods :

For hygroscopic seeding 2 methods are used for enhancing collision - coalescence mechanism.

1. The first method of salt seeding involves application of hundreds of kilograms of salt particles of 10 microns to 30 microns in diameter near the cloud base to produce drizzle sized drops immediately.

2. The second method uses salt flares to disperse about 0.5 to 1 micron size particles into cloud updrafts by releasing one kilogram sized flares from the aircraft from below the cloud base.  Many flares are sometimes released per cloud. Coalescence is enhanced and the seeding material eventually spreads throughout the cloud and if the updraft extends high into the sky then the ice process also gets enhanced. Such hygroscopic seeding is used only on continental type clouds because the maritime clouds already contain many hygroscopic embryos.

Sizes of seeding nuclei :

While investigating whether a large water drop in a cloud will collide with another nearby small droplet, it was found that no collisions occur if the drop diameter is less than the Hocking limit of 38 microns in diameter (the so called Hocking limit of 19 microns is the radius of the drop). If the drop size is less than 20 microns any rain by coalescence may not occur in less than 40 to 60 minutes. Langmuir stated that large raindrops of 5 mm diameter become hydro-dynamically unstable and break into smaller drops which serve as raindrop embryos(about 250 microns) to hasten conversion of many other small cloud droplets into raindrops, particularly in clouds with updrafts above 6 to 8 meters per second strong enough to support such rain drops.

Theoretically for changing the micro-physics of a cloud to produce more rain one may add artificial seeding agents (SCCN) of sufficient size and adequate in quantity to prevent the activation of natural dust particles (CCN) so that the moisture in the cloud gets precipitated.

By introducing particles of 1 to 3 microns at the rate of 25 to 100 per cubic centimeter it may be possible to capture almost all the moisture and prevent the innumerable natural dust particles (CCN) from participating in the cloud formation process. To implement this concept, if 2 micron size sodium chloride particles are used at 50 per cubic centimeter, a cumulus cloud by ingesting at the rate of million cubic meters of air per second, requires seeding rate of 30kg of chemical per minute and this heavy operation is almost impossible to execute.

But if one is prepared to assume that precipitation acts like an infection one can realize that once precipitation starts anywhere in the cloud, it spreads as raindrops and the fragments are carried about by the clouds internal circulations and by the turbulence. Hence a single shot seeding of million cubic meters of air, for example might work. But there is another method of promoting coalescence by the introduction of artificial rain-drop embryos into a cloud by spraying big water droplets into the cloud. Instead of changing all the cloud droplets in the cloud to accelerate coalescence process one can avoid the initial stages of changes in the cloud droplets we are introducing large particles that work as raindrop embryos straight away. But this involves use of very large quantities of water to be sprinkled from an aeroplane and this is very costly. In order to reduce this logistical problem one can treat the cloud with hygroscopic agents like calcium chloride or sodium chloride in the form of dry particles or spray droplets which form rain drop embryos by their own hygroscopic action. This procedure is similar to providing the giant CCN that influence the formation of some of the rain showers over the oceans. However if these artificial rain drop embryos are to be effective the giant hygroscopic particles must be some tens of micrometers in diameter and the particles must produce embryos beyond the Hocking limit of 38 microns in diameter. Moreover the seeding agent must start working as rain drop embryo immediately after getting injected into the cloud. One dimensional cloud models suggest that seeding the clouds with sodium chloride particles of over 120 micron diameter produces rain 10 to 12 minutes after cloud-base seeding.

 But if smaller particles are used the solution droplets not only will grow very slowly but will also often get ejected from the cloud top without ever growing large enough to start falling back down against the updraft currents. Research studies show that hygroscopic seeding is good for clouds with base temperatures above 0oC in general and 10oC in particular. The sizes for the hygroscopic nuclei increase in direct proportion to the updraft speed. For a 5 km deep cumulus cloud with a moderate updraft of 12 meters per second ( 2,400ft per minute) Sodium chloride particles injected into the cloud base must be over 40 microns in diameter for the resultant droplets to avoid ejection from the top of the cloud. However the use of hygroscopic nuclei of 50-100 micron diameter poses a serious logistical problem. If the embryo concentration needed to promote coalescence is estimated at 1000 per cubic meter of cloud, the number of hygroscopic particles to seed the entire cloud will be 1015 and if the particle diameter is 100 microns a few tonnes of seeding material will be required for the cloud.

However if this problem has to be overcome one has to assume that the raindrops formed around the artificial embryos would break into fragments which again create more raindrop embryos. For this purpose raindrops of 2 to 3 mm need to be produced and retarded from coming down through the cloud to experience collisions and break up and this requires updrafts exceeding 5 meters per second (1000 ft per minute) Each cycle of growth and break up and again growth in the cloud takes about 4 minutes. It is not very clear whether these recycling processes would infect the whole cloud before the natural precipitation processes would obtain the same result. Ultimately one has to decide upon a reasonable compromise which requires the use of 25 kg to 50 kg of sodium chloride or other hygroscopic powder like calcium chloride for a cloud of moderate size. This dosage has been found to be reasonable on the basis of results from various experiments conducted in different places. It is necessary to remember that once precipitation appeared anywhere in the cloud the raindrops and raindrop fragments would be circulated throughout the cloud by its organized internal motions and by turbulence. Hence the dosages of chemicals to be used for cloud seeding must be decided on the basis of cloud systems, updrafts, topographical, meteorological parameters etc.

Successful experiments :

1. The experiments conducted at Baramati in India for 11years indicated the success of warm cloud seeding operations by producing 24% additional annual rainfall, significant at a 4% level at a very inexpensive cost with a benefit to cost ratio of 60:1. The Indian Institute of Tropical Meteorology (IITM), Pune launched a warm cloud seeding experiment using aircraft in the semi-arid region towards East of Pune on the lee-ward side of the Western ghats from 1973 to 1986. A randomized double-area crossover design with a buffer zone was used for the aerial seeding work. The experimental area covering 4800 sq. km. was divided into 3 parts designated as North, Buffer and South sectors in Ahmadnagar, Baramati area . The results of experiments on aerial seeding of warm clouds with common salt mixed with a little soap stone powder for the eleven years are presented in the following table. (Table-2)

The results clearly demonstrate that warm cloud seeding has enhanced the rainfall by about 24% on the basis of experiments conducted for 160 days during the eleven summer monsoon seasons.

The physical changes that occurred in the clouds due to seeding with common salt were also recorded with scientific instruments fitted to the air-craft for more than 100 pairs of seeded (target) and not-seeded (control) clouds to provide scientific evidence as positive proof for the efficiency of cloud seeding technique. The physical changes in the seeded clouds are presented here. (Table-3)


The above changes in the micro-physical, dynamical, Electrical characteristics of clouds before seeding and after-seeding along with changes in the concentrations of chlorides and sodium ions in the cloud water and the rain water amply prove that warm-seeding of clouds was mainly responsible for increasing the rainfall by about 24%. (Fig-2)

The experiments demonstrated that shallow clouds with vertical thickness of less than 1 km and liquid water content of less than 0.5 grams per cubic meter showed a tendency for dissipation when common salt was sprinkled over the warm clouds from an aeroplane. The clouds developed rain in 20 minutes following seeding. The giant condensation nuclei (SCCN) of common salt powder used for the experiment had a diameter greater than 10 microns. SCCN concentration in the target cloud was higher by about 2 particles per litre. The salt particles in the seeded clouds had a concentration of 1 per litre of cloud air. The median salt particle mass used for seeding is approximately 10-9 g corresponding to dry particle diameter of 10 micro meters. The seeding material was released into the clouds during the aircraft penetrations at a height of 200-300 meters above the cloud base. The rate of seeding varied between 0 and 30kg per 3 km aircraft flight path on the basis of density of clouds and their vertical thickness in the target area. On the days of theexperiment about 1000 kg of the seeding material was released into the clouds at a slow rate of 10 kg per 3 km for flight path so as to treat as many clouds as possible. Concentration of salt particles released into the clouds were estimated to be 1 to 10 per litre of the cloud air.

2. Some of the warm cloud seeding experiments in South Africa (Mather, 1987) and Thailand (Silverman, 1999) indicated increases in radar estimated precipitation from 30% to 60% from the seeded clouds. Even numerical calculations on the growth of hygroscopic particles into raindrops supported these results on augmentation of precipitation (Cooper 1997).

3. South Dakota experiments from 41 salt seeded and 38 unseeded cases indicate that first radar echoes appeared closer to cloud base in salt seeded than in unseeded clouds, the average height of the echoes being 1.5km for salt seeded and 3km for unseeded clouds above their bases. Rainfall can be increased from a cloud of about 5km depth by salt seeding but the effect decreases with increasing cloud size.

Points to remember :

1. Hygroscopic substances are those which attract water or encourage condensation of water vapour into liquid water upon themselves at relative humidities less than 100%. Introduction of artificial embryos into warm clouds include the use of water sprays, hygroscopic powders and hygroscopic solutions.

2. The cloud drops of 5 microns to 20 microns diameter must grow to precipitation size of over 500 microns by coalescence in cumulus clouds, based upon updraft velocities and sizes, water content, lifetime of the cloud and the initial cloud drop size distributions which again are controlled by the size spectrum of condensation nuclei (CCN).

3. Over the oceans the giant sea salt nuclei of 1 micron to 10 microns in radius become the major factors in initiating coalescence, making marine cumulus clouds give rain more readily than their counterpart continental cumulus clouds over the land.

4. According to Langmuir Chain reaction the cloud drops grow by coalescence into raindrops of 6 mm size with velocity of 10 meters per second which break into fragments that again are carried up by the updrafts and again grow to rain drop sizes and fall on earth as rain.

5. Hygroscopic materials of 1kg may result in the collection of 5 to 10 kg of cloud water into artificial rain drop embryos which coalesce with other droplets and give substantial rainfall.

6. When warm cloud is seeded with NaCl at its base to form artificial raindrop embryos in concentrations of 103/m3 the temperature increases by about 0.05oC.

7. Introduction of artificial embryos of 20 microns to 100 microns in diameter into the lower part of a continental cumulus cloud may result in formation of rain 10 to 15 minutes before natural rain could form.

8. It is said that seeding agents of about 100 kg per warm cloud are required unless the  chain reaction can be initiated. It means that clouds with updrafts strong enough to support the Langmuir Chain reaction may be taken up for experiments by specifying the conditions favouring chain reaction growth.

9. But modelling studies and field experiments show that dynamic effects of hygroscopic seeding

cannot be undertaken intelligently in any programme of cloud seeding operations.

Determining sizes of hygroscopic particles :

Systematic modeling was done to map the growth and trajectory of hygroscopic particles by using particles from 5 microns to 40 microns in diameter, updrafts from 1 to 25 meters per second with cloud base heights of 1 to 3 km and base temperatures upto 20oC. The study was conducted by the experts of the United States Bureau of Reclamation, G.E.Klazura and C.J.Todd (1978). The case study of salt seeding by Biswas and Dennis of July 1970 was also analysed by using this model. This study used a 1-dimensional condensation-coalescence model that follows the growth of a single precipitation particle in a given cloud environment. The process simulated involve release of hygroscopic particles in the updraft region below the cloud and the particles grow by condensation only until they penetrate the cloud base. Later both condensation and coalescence occur until the particle grows to 100 micron diameter at which point growth starts by only coalescence. A drop-frezing temperature of -15oC was selected. When a drop grows to 5mm diameter it breaks into a 2.5mm drop and small fragments. Growth continues only on the 2.5mm drop and not on the fragments. Input to the model are cloud base height, relative humidity, liquid water content and updraft profiles, soundings (height, pressure,mixing ratio and temperature) and initial sizes and physical characteristics of particles like molecular weight, density and van’t Hoff factor. The model output presents both trajectories and growth patterns for different size particles. In the graphs, the trajectories of particles with height and time and the liquid water content profile and curves of dotted or solid lines used for different size particles. The curve labeled UPD indicates trajectory of a parcel that would ascend exactly at the speed of the updraft.

Multiple computer runs were made for cloud conditions, characteristic of temperature and warm regions. Relative humidity was 100% at cloud base with a super saturation of 0.1percent at 100 meters above cloud base upto the top. 5 to 400 micron particles were introduced at 0.5km for cloud bases of 1,2 and 3km and in-cloud saturation adiabats (qm) of 16 and 23.2oC A 5 micron particle with updraft of 2 meters per second rises to about 2km height above the base grows to 2 to 4 mm diameter and falls from the base of the cloud in 24 minutes while a 40 micron particle goes upto about 0.5 to 1km in height and touches the base in 18 minutes.  (Fig-3)

For an updraft of 5 meters per second the 5 micron particle goes upto about 4kms in the cloud and breaks 5 times and the rain drops reach the base in about 20 minutes. In the case of an updraft of 10 meters per second the 5 micron particle gets into the freezing zone and Langmuir chain reactions occur when the drop attains 5mm size and breaks up into fragments which are pushed up the updrafts and grow again to rain drop sizes and fall on earth.. But for updraft of 10meters per second, the 5 micron particle would rise high to freeze before reaching 5mm size and therefore no breakup occurs in it. However the large drops of 10 to 400 microns get into the accumulation zone and migrate towards the periphery of the updraft core and as the updraft becomes less vigorous the drops fall down. In the case of stronger updrafts the smaller sized particles are all ejected out of the top of the cloud while the large 100 microns and 200 microns grow large enough to fall back through the updrafts as hail stones. Hence parameters like cloud base height and temperature and cloud thickness and diameter, updraft and cloud water contents and cloud droplet-spectra influence the efficiency with which hygroscopic particles act as cloud seeding material to stimulate additional precipitation from a cloud

The numerical simulations demonstrated that for slow updrafts the larger hygroscopic seeding agents travel through the lower trajectories and sweep out lesser water than the small hygroscopic particles grow into raindrops and break up to initiate Langmuir chain reaction while the smaller hygroscopic particles are carried upwards to the cirrus level and get lost before they attain the precipitation sizes. In case of very strong updrafts only the large hygroscopic particles get a chance to reach precipitation size and in this situation hail stones are produced.

The warm cloud seeding experiments are based on static mode seeding concept or seeding for microphysical effects whereby the seeding agents mainly work to improve efficiency of rain formation by accelerating the trinity phenomena of condensation-coalescence-collision process in the cloud. The Indian and Thailand experiments injected into the cloud giant hygroscopic giant CCN of over 10 microns in diameter mainly to jump-start the collision coalescence process by what is known as “Brute Force” seeding approach which needs very large aircraft that can carry tonnes and tonnes of seeding chemicals. In this process the salt particles are so big that they produce comparatively few rain drops for the large seeding mass employed. Even then the seeding process apparently has produced the desired results in India and Thailand. But the hygroscopic flares based on 0.5 micron diameter CCN aerosols affect the condensation process substantially by broadening the initial cloud drop sizes that promote thecompetition affect whereby the larger nuclei are activated preferentially over the smaller nuclei. The flare particles of less than one micron diameter are found by the scientists to have a negative effect on the rain development. Flare technology has 2 major drawbacks Firstly, the hygroscopic particles produced by the flares being less than one micron in diameter are smaller than the optimal size according to the model studies. Secondly, the quantity of actual material used for seeding is smaller than is required as per the hypothesis of hygroscopic seeding for microphysical effects. But if more flares are needed then it is going to increase the cost of the operations. In order to overcome the short-comings of the hygroscopic flare seeding technology, the experts have to fine tune the flares to produce fewer number of small particles and more number of optimum size particles which should be large than one micron in diameter. Another approach for an alternative flare technology is to produce more hygroscopic CCN particles in the optimum range by using hygroscopic powder sprays into the clouds. Model simulations by scientists indicated the need for hygroscopic particles of 3 to 5 microns diameter and for one of the experiments such salt particles were dispersed into warm clouds from an agricultural sprayer or a duster aircraft. In one of the experiments in Texas about 20kg of salt powder was used by the duster aircraft for about 7 minutes to seed the cloud at the seeding rate of 3kg per minute and this is a higher seeding rate than the 0.5kg per minute seeding rate typically used with flare seeding methodologies. The experiments indicate microphysical changes in the clouds which are primarily due to the effect of the giant salt nuclei and secondarily through a week competition effect. Some scientists consider that the proper sizes for hygroscopic aerosols are between 2 to 5 microns in diameter. However more research work has to be done to determine the optimum sizes of hygroscopic particles for warm seeding based upon geographical, meteorological and topographical features of a given region.