Stormfury Cumulus Seeding Experiment 1965: Statistical Analysis and Main Results

View More View Less
  • 1 Environmental Science Services Administration, Washington, D.C.
© Get Permissions
Full access

Abstract

A randomized seeding experiment was carried out on 23 tropical oceanic cumulus clouds on 9 days in the summer of 1965 as part of the joint Navy-ESSA Project Stormfury. Following instructions in sealed envelopes, an aircraft seeded 14 of the clouds with 8-16 pyrotechnic silver iodide generators called Alecto units. Each unit releases about 1.2 kg of silver-iodide smoke. The nine remaining clouds were studied in an identical manner as controls, using the same stack of four instrumented aircraft to penetrate the cloud before and after the seeding run. Cloud growth was documented by aircraft, radar and photogrammetry. The seeded clouds grew vertically an average of 1.6 km more following the seeding run than did the control clouds; the difference is significant at the 0.01 level.

A numerical model of cumulus dynamics was specified in advance of the field program. This model integrates the equation for the vertical acceleration of an entraining cumulus tower, predicting top heights of unseeded and seeded clouds as a function of ambient sounding and horizontal tower dimension. Seedability is defined as the predicted difference between the seeded and unseeded top of the same cloud. Effect of seeding is defined as the difference between the observed top and the predicted unseeded top of the same cloud. Both parameters are computed and graphed for all 23 clouds. Seeded and unseeded clouds separate into distinct populations. This statistical analysis demonstrates that 1) seeding has a clear effect on cumulus growth under specifiable conditions and 2) the model has considerable skill in predicting the amount of growth and in specifying the conditions.

Sources of subjectivity and bias are shown to be small and not to affect the results. The sensitivity of the model predictions to variations in input data is investigated with two examples, one each of large and of negligible cloud growth following seeding. Some possible effects of natural glaciation are examined with the model and future phases of the program are described.

Abstract

A randomized seeding experiment was carried out on 23 tropical oceanic cumulus clouds on 9 days in the summer of 1965 as part of the joint Navy-ESSA Project Stormfury. Following instructions in sealed envelopes, an aircraft seeded 14 of the clouds with 8-16 pyrotechnic silver iodide generators called Alecto units. Each unit releases about 1.2 kg of silver-iodide smoke. The nine remaining clouds were studied in an identical manner as controls, using the same stack of four instrumented aircraft to penetrate the cloud before and after the seeding run. Cloud growth was documented by aircraft, radar and photogrammetry. The seeded clouds grew vertically an average of 1.6 km more following the seeding run than did the control clouds; the difference is significant at the 0.01 level.

A numerical model of cumulus dynamics was specified in advance of the field program. This model integrates the equation for the vertical acceleration of an entraining cumulus tower, predicting top heights of unseeded and seeded clouds as a function of ambient sounding and horizontal tower dimension. Seedability is defined as the predicted difference between the seeded and unseeded top of the same cloud. Effect of seeding is defined as the difference between the observed top and the predicted unseeded top of the same cloud. Both parameters are computed and graphed for all 23 clouds. Seeded and unseeded clouds separate into distinct populations. This statistical analysis demonstrates that 1) seeding has a clear effect on cumulus growth under specifiable conditions and 2) the model has considerable skill in predicting the amount of growth and in specifying the conditions.

Sources of subjectivity and bias are shown to be small and not to affect the results. The sensitivity of the model predictions to variations in input data is investigated with two examples, one each of large and of negligible cloud growth following seeding. Some possible effects of natural glaciation are examined with the model and future phases of the program are described.

Save