WMOD00439

• Title: HIPLEX-1: Experimental Design and Response Variables
• Date: 4/4/1984
• Weather Modification - Doc Type: Report

498

JOURN AL OF CLIMATE AND APPLIED METEOROLOGY

VOLUME 23

temperatures as low as -160C(McPartland et aI., 1977;
Cooper, 1978a). The available evidence (see also Cerni,
1982, and Lamb and Hallett, 1982) suggested that, at
least in most small clouds, ice multiplication processes
were relatively unimportant. This combination of con-
ditions indicated a possible role for glaciogenic seeding
agents.
A scientific. workshop convened at Dillon, Colorado,
in December 1977 (Bureau of Reclamation, 1977) to
deliberate upon the key scientific issues and attempt
to outline a specific experimental design. The workshop
identified a number of issues that remained to be re-
solved, but after another summer of field study and
some additional analysis, the HIPLEX scientists felt
ready to undertake a randomized seeding experiment
on single clouds. The complexity of the precipitation
processes was such that it was agreed that the Phase
2 experiments should progress from relatively simple
cumulus congestus clouds in HIPLEX-l to larger and
more complex clouds and mesoscale systems in sub-
sequent experiments. The following quotation from
the preface of the HIPLEX-l design document (Bureau
of Reclamation, 1979) helps set the stage:

HIPLEX-I is the first in a series of experiments under the
High Plains Cooperative Program aimed at establishing
the physical basis for enhancing beneficial growing-season
precipitation from convective clouds on the High Plains.
The experimental design is based on three years of research
on the physics of natural clouds in eastern Montana. . . .
The experiment is focused on cumulus congestus clouds
and explores the hypothesis that both the probability and
amount of rainfall will be increased by means of ice-embryo
seeding for microphysical effects. The experiment is one
in which each step leading to the additional precipitation
is specified in advance and verified by observations during
the course of the experiment.

The three years of exploratory study included in-
vestigations of cloud and precipitation climatology;
aircraft and radar studies of cloud physics processes;
and the development and testing of the equipment and
techniques needed for the contemplated seeding ex-
periment. The results of those studies provided the
essential background to complete the detailed exper-
imental design, which was developed in early 1979
(Bureau of Reclamation, 1979). The purposes of this
paper are to summarize the design and conduct of
HIPLEX-l; to present the measured values of the re-
sponse vanables; and thereby to set the stage for two
companion papers in this issue (Mielke et al., 1984;
Cooper and Lawson, 1984) which discuSs the statistical
and physical evaluations of the experiment.

2. Overview of the design of HIPLEX-l

HIPLEX-l was viewed as an exploratory experiment
representing the initial part of Phase 2, the single-cloud
rain modification experiments. The core of HIPLEX-
I was a randomized seeding experiment on individual,

small, semi-isolated, cumulus congestus clouds (Sil-
verman, 1980). The experiment was similar in some
respects to that conducted by Bethwaite et al. (1966),
but for HIPLEX-l a much greater array of instru-
mentation was available. The seeding method involved
dJ;opping dry ice pellets into the clouds in an effort to
produce "static" or microphysical effects, namely the
early formation of precipitation-sized ice particles in
them. In these clouds, which have short lifetimes, the
resulting acceleration of the precipitation process was
expected to increase both the fraction of the clouds
that precipitated and the amount of precipitation. In
view of the environmental conditions, the micro-
physical characteristics of the clouds for which the
experiment was designed, and the light seeding rates
used, any ancillary dynamic effects were expected to
be negligible (Cooper, 1978a). Parallel studies of eco-
nomic, environmental, and sociopolitical aspects of
the program were also conducted, but they will not be
dealt with here.
The design of the experiment (Bureau of Recla-
mation, 1979) comprised several major components,
including the following:

1) Definition of the experimental unit and a detailed
procedure for identifying suitable clouds.
2) A detailed experimental procedure for treating
the selected clouds, with random allocation between
seed and no-seed treatments, and for measuring their
responses to that treatment.
3) A multistep physical hypothesis outlining the se-
quence of anticipated effects of the seeding.
4) A set 'ofresponse variables established to monitor
the developments corresponding to each step of the
physical hypothesis.
5) A set of statistical hypotheses associated with the
expected behavior of the respective response variables.
6) A plan for statistical evaluation of the results of
the experiment.
7) A plan for physical evaluation of the experiment.

Subsequent sections of this paper discuss the first five
of these components. The sixth is discussed in the
companion paper on the statistical evaluation (Mielke
et al., 1984), and the last in the companion paper on
the physical evaluation (Cooper and Lawson, 1984).
Particularly for items 1-4 in the foregoing list, the
development of the final experimental design repre-
sented a synthesis between the expected behavior of
the clouds and the capabilities of the instrumentation
available. Thus the design underwent several iterations
to assure that the anticipated responses were specified
in terms of quantities measurable with the available
equipment. In this process, numerous scenarios were
postulated to test the robustness of the design; if a
scenario could be devised that would lead to an er-
roneous inference from some aspect of the experiment,
the design was modified to prevent that. An example
in Section 5 illustrates this approach. It is, of course,