<br />."
<br />
<br />(1975a), Gagin et al. (1986) found no
<br />difference in the relationships between
<br />cell height and other variables,
<br />including rainfall volume, for unseeded
<br />and silver iodide seeded cells.
<br />Therefore, they interpreted all of the
<br />apparent FACE seeding effect in terms of
<br />changes in cell height.
<br />
<br />On the other hand, Rosenfeld and
<br />Woodley (1989) found for a seeding
<br />experiment in Texas a "large indicated
<br />effect of treatment on cell rain volume,
<br />despite the small indicated effect on
<br />maximum cell heights." It is
<br />interesting to note that the Texas
<br />result, which is only tentative due to
<br />the small number of cases analyzed to
<br />date, is in line with the Cloud Catcher
<br />seeding hypothesis, which stressed a
<br />combination of microphysical and dynamic
<br />effects, rather than the hypothesis in
<br />the Texas project design, which was
<br />patterned after FACE and considered
<br />increases in cell height as the
<br />principal, if not the sole, source of
<br />additional rainfall (Rosenfeld and
<br />Woodley, 1989).
<br />
<br />Any future experiments near the
<br />Black Hills should be designed to shed
<br />light on these questions, as well as
<br />interactions between the clouds and the
<br />Black Hills themselves. Seeding clouds
<br />over the Black Hills to maximize
<br />precipitation requires consideration of
<br />wind, temperature, and moisture fields
<br />over the entire area. Kuo and Orville
<br />(1973) showed how the interactions
<br />between the Black Hills and the air
<br />streams passing over them set up
<br />preferred areas for cloud and shower
<br />formation. Because seeding influences
<br />cloud dynamics, one should not assume
<br />that seeding over one part of the Black
<br />Hills does not influence subsequent
<br />cloud developments elsewhere.
<br />
<br />A unifying concept, which seems to
<br />bring together various pieces of
<br />evidence, is that rainfall increases are
<br />produced when seeding helps convective
<br />clouds merge into clusters, rather than
<br />continuing as single-cell showers
<br />(Dennis et al., 1976). Such
<br />organization promotes the precipitation
<br />efficiency of clouds. It also appears
<br />to increase the total amount of water
<br />vapor processed, which accounts for the
<br />great attention paid to cloud mergers in
<br />cloud seeding experiments in Florida and
<br />Texas (e.g., simpson and Dennis, 1974;
<br />Rosenfeld and Woodley, 1989).
<br />
<br />The apparent success of seeding on
<br />Rapid Project shower days has already
<br />been noted, as well as the lack of
<br />overall success on storm days. On Rapid
<br />
<br />Prc)j ect storm days with southwesterly
<br />flow, there were indications of rainfall
<br />decreases at 1:he edge of the Black
<br />Hills, but of substantial inc::reases some
<br />30 to 40 km northeastward, still within
<br />the target areas (Dennis et al., 1976).
<br />It appears that, for that particular
<br />situation, additional organi:~ation or
<br />clustering induced by seeding was still
<br />beneficial over the open prairie, but
<br />could not be produced (or was not
<br />beneficial) over the Black Hills
<br />the!msel ves .
<br />
<br />These inferences are only tentative,'
<br />but: suggest a possible stratE!gy to be
<br />followed in any future field
<br />experiments. On shower days, only
<br />clouds supported by low level
<br />convergence induced by the Black Hills
<br />themselves have any chance of producing
<br />significant showers. Seedinsr would be
<br />directed at the most promising growing
<br />clouds, wherever they occurre!d. On
<br />storm days, as evidenced by positive
<br />vorticity advection (synoptic-scale
<br />forcing), systems over the Black Hills
<br />would organize with no help, and seeding
<br />there likely would have no effect.
<br />However, clouds forming in the less
<br />disturbed regions might be of a size
<br />(COP around 4 to 6 km) that could be
<br />stimulated to further growth and
<br />organization. Evaluation should be on
<br />both a cell or cluster basis and an
<br />area-wide basis to study the larger
<br />scale effects.
<br />
<br />continued work with powerful
<br />numerical models, such as the new three-
<br />dimensional models developed by Clark
<br />and others (e. g., SmolarJdewicz et al.,
<br />1988), is required to sort out the
<br />various possible effects. In time, it
<br />should be possible to integrate detailed
<br />parameterizations of microphysical
<br />processes in individual clouds into
<br />three-dimensional mesoscale models
<br />capable of simulating air motions over
<br />and around the Black Hills. The output
<br />of such models would suggest additional
<br />seeding hypotheses, and might even
<br />provide statis.tical controls for seeding
<br />experiments. This would be an extremely
<br />important development, as it 'would
<br />provide to single-area experiments a
<br />degree of sensitivity now obtainable
<br />only with target-control or randomized
<br />crossover experiments. As we have seen,
<br />the interpretation of results from
<br />target-control or randomized crossover
<br />experiments is ambiguous when strong
<br />dynamic effects are present or
<br />suspected.
<br />
<br />9
<br />
|