Skip to main content

Cloud seeding

At the cessation of the Second World War Taffy Bowen and his staff at the Division of Radiophysics began to look around for work of interest to themselves and of importance to Australia. One of those areas was cloud and rain physics which Bowen initiated and led until his retirement in 1971. For over five decades, the study of clouds, rain and the atmosphere has been the work of the CSIRO Division of Cloud Physics, now known as Marine and Atmospheric Research.

The ‘rainmakers’ carried out experiments over South Australia, Tasmania, the Snowy Mountains, the Warragamba Dam catchment area west of Sydney and New England. They used dry ice and silver iodide to ‘seed’ clouds, which resulted in the production of rain.

The first year’s results were tremendously heartening, with rainfall increases of up to 30 per cent in the target areas. But frustration followed. Rainfall appeared to deteriorate and was more variable in the target areas than before the experiments started.

Today, CSIRO uses the lessons learnt from their cloud seeding experiments to develop better models for weather forecasting and changes in climate.

Early success

In 1946 USA researchers I Langmuir and V Schaefer reported that rain could be induced by seeding clouds with dry ice. While many reacted cautiously to these claims Bowen immediately saw the potential importance of the technique for dry Australia. Within months, two members of his staff had investigated Langmuir and Schaefer’s work and, on their return, had carried out a trial in eastern New South Wales using RAAF aircraft. Success was immediate – the date was 5 February 1947. It was a day when deep cumulus cloud covered the country inland from Sydney. All the clouds appeared similar in type and size which was important for a clear-cut result. A plane dumped dry ice into one cloud and within minutes rain started to fall while the cloud-top mushroomed explosively. The rain lasted several hours and more than 12 millimetres fell over an area of 80 square kilometers. Surrounding clouds gave no rain. This is believed to be the first documented case anywhere in the world of an appreciable man-made rainfall reaching the ground and the first time that dynamic cloud growth had followed seeding. This striking result held such promise that a systematic program of cloud seeding was set up in February 1947 and continued for the next twenty-four years.

In Australia, where fickle rainfall has elated and then downcast countrymen from the time the first pioneers saw once brimming rivers and lush pastures fade to muddied waterholes and dustry earh, it was almost inevitable that the weather modification work of the Division of Cloud Physics should concentrate on rain-making. The Division’s work included theoretical, laboratory and airborne investigations of cloud structure and reaction.

The challenge

Natural variability of rain has been the rain-makers’ single biggest headache. As the scientists recalled:

It makes it terribly difficult to prove anything. You can go to an area and influence rain-potential clouds so that it looks as if you have increased the rainfall. But how much rain would have fallen if you hadn’t interfered with them? On one occasion you simply can’t tell. But if you keep on repeating the experiment, and keep on increasing the rain, eventually you can prove you caused the increase.

In this cloud seeding research, the Division of Cloud Physics worked closely with State Departments of Agriculture. Each Department had regional referees to evaluate rain needs and public opinion in various areas. Before any rain-making program started, public meetings were held to vote on the issue. But the weather is never right for everyone and the Division received letters from irate landowners blaming rain-making experiments for unwanted downpours.

Multiple approaches

As little was known about the properties of clouds in Australia or the mechanisms of rainfall, Bowen initiated a vigorous research program of cloud studies. This included not only the effects of adding dry ice to cold clouds, but also the effect of spraying water into warm clouds which are responsible for much of the rainfall in the warmer parts of Australia. Bowen took part in the latter work himself and during 1950-55 published papers on the theory of coalescent rainfall and directed experimental trials.

Dry ice has a temperature of -80 °C or colder. If a piece the size of a pea is dropped into a supercooled cloud it will fall as far as three kilometres before evaporating completely, leaving a wake of ice crystals. In the right conditions, each crystal will feed on cloud droplets to form a large snowflake which melts to a raindrop as it reaches lower and warmer levels. This attractively simple principle was used from 1947 to 1950 near Sydney when 45-kilogram loads of dry ice were dropped into suitable clouds, their near neighbours being left unseeded to provide a basis for comparison. The principle seemed to work best with continental cumulus cloud masses where the air was dirty so that lots of small droplets were formed which were unlikely to coalesce of their own accord.

The difficulty with this method of stimulating rainfall was that only a few clouds could be treated on any one day and large amounts of dry ice were required. This limitation was overcome by the discovery, again in the USA, that tiny quantities of silver iodide smoke could be used as a seeding agent. Unlike many of his contemporaries, Bowen saw the potential for seeding large areas from the air using silver iodide burners mounted on an aircraft. Silver iodide smoke particles provide ‘kernels’ on which ice crystals can grow in a supercooled cloud. Theoretically, grams of silver iodide will do much the same job as kilograms of dry ice, so that smaller and cheaper aircraft can be used. Silver iodide seems to work best in layer clouds formed in air coming in from the sea.

From 1955 to 1963 the rain-makers carried out four intensive experiments over South Australia, the Snowy Mountains, the Warragamba Dam catchment area west of Sydney, and the New England region of NSW. For each experiment there was a target area of 2 000 to 8 000 square kilometres and a neighbouring control area of the same size which was not seeded. A network of up to 150 rain gauges covered each area. The first two years were so successful, with an estimated rainfall increase of 25%, that several more regions were quickly selected. There the early indications were also successful, but in many subsequent years all areas showed a gradual decay of the induced rainfall with time. Most people would have become discouraged by such a result and given up. Bowen, however, proposed a simple explanation, based on the idea that a persistence phenomenon

The Tasmanian cloud-seeding experiments

Tasmania was chosen for subsequent cloud-seeding trials. The experiments were designed to compensate for quirks in the results of previous experiments, which had frustrated the rain-makers and led them to a serious re-evaluation of their programs. This experiment in 1971 was a success but since Bowen had now retired, the result was not immediately attributed to the correctness of his persistence hypothesis

Subsequent work by EK Bigg has done much to explain the detailed mechanism of the phenomenon. With the continuing success of cloud seeding work by the Australian states of Tasmania and Victoria and the recognition of the role of persistence, there appears now to be a promising future for the rain making techniques that Bowen did so much to pioneer.

Bowen’s theories on periodic rainfall

Bowen’s remarkable energy and enthusiasm were evident also in other programs. He was not afraid to speculate and presented his intuitive ideas with a persuasive and engaging optimism that was either inspiring or alarming to his colleagues, depending on their views of science. Two of his well known theories about periodic rainfall variations illustrate this.

The influence of meteor showers

From the daily rainfall records for Sydney over the period 1859 to 1952 and for stations elsewhere in New South Wales and in other countries, Bowen found well defined peaks of rainfall in January and February. These anomalies he correlated with the passage of the Earth, 30 days earlier, through specific meteor streams that orbit the sun. He suggested that the smaller particles fell through the atmosphere to cloud level in 30 days, where they induced the observed rainfall.

The apparent physical implausibility of this hypothesis attracted a wave of criticism: the number of particles was insufficient, the fall time would not be fixed, and the particles would not form ice crystals. Even the reality of the anomalies was vigorously questioned, but independent analysis showed that they were statistically significant. But Bowen was not impressed by purely statistical arguments and insisted that his staff probe crucial aspects of his hypothesis by empirical tests in clouds. Whether he was right to invoke meteor showers to explain the rainfall anomalies and if so, how they influenced clouds after a fixed time interval, has yet to be demonstrated.

Lunar effects

In 1962, following a paper published in the USA, Bowen and Adderley showed that there were similar lunar effects in the monthly rainfall records for fifty New Zealand stations with comparable magnitude and closely related phase. The reality of the effect was beyond doubt. Independent frequency analysis revealed an amplitude variation of 20% and a periodicity of 29.5307 days. The mean period between full moons is 29.5306 days.

Bowen suggested that the Moon, revolving about the Earth, could modulate the amount of meteor dust reaching the Earth, and later showed that meteor rates in both the northern and southern hemispheres varied similarly with lunar phase. He argued that the Moon could intercept the particles or alternatively could deflect them because of electrostatic charges on the Moon and particles. Modern studies by his colleague, EK Bigg, however, suggest that the Moon’s influence on rainfall is more likely to be caused by the lunar tides in the Earth’s atmosphere.


The cloud and rain physics group, under Bowen’s leadership, worked in a most stimulating environment. Even his more speculative ideas sometimes drove his critics to discover truths that would otherwise have remained hidden. Over twenty-four years, the group established a high international reputation with its achievements and an impressive number of sound scientific publications.