Numerical Simulation of Cloud Seeding Experiments in Selected Australian Areas

S. Twomey Division of Atmospheric Physics, CSIRO, Sydney, Australia

Search for other papers by S. Twomey in
Current site
Google Scholar
PubMed
Close
and
I. Robertson Division of Atmospheric Physics, CSIRO, Sydney, Australia

Search for other papers by I. Robertson in
Current site
Google Scholar
PubMed
Close
Restricted access

Abstract

Using actual rainfall records for selected areas, one can by computer simulation apply hypothetical seeding effects and then compare the results obtained and inferences drawn by individual “experimenters” who have superimposed the effects of seeding on the natural variations of rainfall in space and in time. In this way it is possible to infer which kinds of experimental design are preferable and the duration of experiment likely to give meaningful results in a given area for some prescribed increase by seeding. Some results for three different Australian areas are given. These show a marked superiority of the crossover design as compared with target-control or single-area experiments. They also suggest that in more favorable areas a meaningful result could be obtained with a four- to eight-year experiment when the average increase due to seeding was 10%, but in the less favorable (arid) areas even a 16-year experiment would not be long enough. An average 20% increase would, however, be detectable with confidence even in the least favorable area but it would require an experimental period of about 16 years.

Abstract

Using actual rainfall records for selected areas, one can by computer simulation apply hypothetical seeding effects and then compare the results obtained and inferences drawn by individual “experimenters” who have superimposed the effects of seeding on the natural variations of rainfall in space and in time. In this way it is possible to infer which kinds of experimental design are preferable and the duration of experiment likely to give meaningful results in a given area for some prescribed increase by seeding. Some results for three different Australian areas are given. These show a marked superiority of the crossover design as compared with target-control or single-area experiments. They also suggest that in more favorable areas a meaningful result could be obtained with a four- to eight-year experiment when the average increase due to seeding was 10%, but in the less favorable (arid) areas even a 16-year experiment would not be long enough. An average 20% increase would, however, be detectable with confidence even in the least favorable area but it would require an experimental period of about 16 years.

Save