WMOD00572

• Title: Status of Precipitation Augmentation and Hail Suppression Experiments
• Date: 2/20/1990
• Weather Modification - Doc Type: Report

Printed January 30, 1990

HIPLEX-l clearly showed that dry ice seeding produces ice crystals in supercooled cumulus clouds,
and that the crystals grow and spread at rates agreeable with theory (slides) (Mielke et aI., 1984;
Cooper and Lawson, 1984). The dry ice drops produced curtains of tiny ice crystals, which spread
laterally at roughly 1 m.s'l on each side. In some cases the crystals grew into graupel. However,
cloud lifetime was a limiting factor in most of the test clouds. All of the liquid water concentrations
measured more than 5 minutes after seeding were far below those required for the development
of millimeter-size raindrops, and the precipitation that formed took the form of snow aggregates
rather than graupel (Cooper and Lawson, 1984). Most of the test clouds obviously failed to satisfy
criterion (c) for seed ability stated by the WMO.

Similar findings have been reported from seeding experiments on convective clouds near Thunder
Bay, Ontario and near Yellowknife in the Northwest Territories of Canada (Isaac et aI., 1982). In
those experiments, silver iodide was released from wing-mounted pyrotechnic flares as the seeding
aircraft penetrated the clouds at some level between -5 and _lOOC. About 40 percent of the
Yellowknife seeded clouds rained compared to only 12 percent of the unseeded clouds. However,
no rain was generated from the Thunder Bay clouds. The difference apparently was related to cloud
lifetimes after selection as test cases, where lifetime is defined as the time that the peak liquid
water concentration remained above 0.1 g.m'3. Many of the Yellowknife clouds had lifetimes of
15 to 30 minutes, and most of the long-lived ones rained. Only two of the Thunder Bay clouds had
lifetimes over 10 minutes, and many dissipated in less than 5 minutes, which is far too short a time
for precipitation to form.

A logical next step in research on convective cloud seeding would be an experiment in which the
capabilities developed for HIPLEX-1 and the latest numerical cloud models would be directed at
clouds large enough to yield appreciable rainfall at the ground. Convective cloud seeding programs
have been conducted at various places in Texas throughout the 1980s, including some randomized
trials supported by the Bureau of Reclamation. Preliminary results as reported by Rosenfeld and
Woodley (1989) have been quite promising, with indications of a doubling of rainfall from seeded
clouds. However, only a few test cases have been analyzed so far. The Illinois State Water Survey
has made detailed plans for a more comprehensive study, but field activity has been limited by a
lack of funding.

5. Hail Suppression

Hail is formed in large convective storms. They are generally classified as single-cell, rnulticell, or
supercelI storms. It is convenient to divide hailstone growth into two stages. The first is the growth
of a hail embryo about 5 mm in diameter; the second is the continued growth of the embryo to a
size, say 2 or 3 cm diameter, large enough to reach the ground without melting completely. Analysis
of hailstone trajectories suggests that embryo growth takes place in updrafts of 1 m.s.l or so, while
final growth may take place in updrafts of 20 or 30 m.s'l. Such a combination could be achieved
in a single-cell cloud with a steadily intensifying updraft, by recycling embryos in a multicell cloud,
or by trajectories that carry hail embryos from the outer parts of a quasisteady-state supercell cloud
inward toward the main updraft (Foote and Knight, 1977).

A number of conceptual models have been put forth to suggest how silver iodide seeding could
reduce hail damage; of these, the competing embryo hypothesis is considered the most promising

9