180 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1951 



provide buoyancy forces in the lower levels. Air cannot be "sucked" 

 up but must be driven up by a force applied to each element of the 

 air column. There will be a tendency for the supercooled portion of 

 the cloud to separate from the portion below the freezing level. Un- 

 less the environment above this level is very moist and the lapse rate 

 steep, it is doubtful that there will be any marked growth of the upper 

 portion of the cloud. If there is to be an acceleration of the cloud 

 development as a result of the heat released by seeding, a natural 

 convective circulation must be active in the lower layers. If this 

 natural convection has been terminated somewhere above the freezing 

 level by a small inversion or a slightly stable lapse rate, the heat re- 

 leased by seeding may extend the convective activity upward. This 

 requires rather special conditions. It may be remarked that a tempera- 

 ture rise of 1° C. in the upper portion of a cloud is not very impressive 

 when viewed on an adiabatic chart. The evaporation of the solid 

 carbon-dioxide pellets will cause the cooling of the air through which 

 they pass. This will cause localized subsidence, but in view of the 

 small amounts of carbon dioxide used this cooling will have little 

 effect after the first few seconds. 



An effect similar to the heating discussed above will result if the 

 liquid water of the supercooled cloud is caused to precipitate. This 

 removal of mass will result in an upward accleration. It is easy to 

 show that the removal of 1 g of water per cubic meter will cause the 

 same vertical acceleration as a temperature rise of about 0.3° C. Note, 

 however, that the falling precipitation will impose a downward accel- 

 eration on the lower levels of the cloud. This downward acceleration 

 will increase as the precipitation elements grow in their fall through 

 the cloud. It is believed that this downward force of the precipitation 

 is partly responsible for the downdraft in a mature thunderstorm. 

 The convective circulation in a thunderstorm is strong enough to main- 

 tain an upward motion in part of the cloud in spite of the downdraft. 

 It is hard to see how the release of precipitation from a seeded cloud 

 can increase the convective circulation. 



It is reasonable to conclude that seeding only rarely will stimulate 

 the vertical development of a cloud. In the vast majority of cases dis- 

 sipative effects are to be expected. These conclusions seem to be borne 

 out by the results of cloud-seeding tests. 



ARTIFICIAL VERSUS NATURAL NUCLEI 



The basic premise of the method for the release of precipitation by 

 seeding is that there is often an insufficient number of natural nuclei of 

 sublimation or crystallization present in supercooled clouds to initiate 

 precipitation by the Bergeron-Findeisen mechanism. The regular 

 existence of supercooled clouds in the atmosphere supports this as- 



