WMOD00476

• Project Name: Colorado River Basin Pilot Project
• Title: Executive Summary of Comprehensive Evaluation
• Date: 12/1/1976
• Weather Modification - Doc Type: Report

2. ANALYSIS AND EV Al"UATION

The comprehensive evaluation was organized so that the basic physical
factors influencing seeding effectiveness, such as the supply of cloud water,
the supply of natural nuclei, the trajectory of falling snow, and the degree
of thermodynamic instability of the clouds, were represented in terms of
parameters based upon well observed variables. This enabled evaluators to
examine the role played by these factors and develop a method for transferring
this knowledge to mountain barriers having different cloud climatologies and
different terrain features.

t
]

The evaluation took two forms: 1) a formal a priori statistical analy-
sis based on the precipitation data for 71 experimental treated days and 76
experimental control days, and 2) an a posteriori statistical analysis of
precipitation and upwind sounding data arranged in six-hour blocks.

The grouping into six-hour blocks, rather than three-hour blocks asso-
ciated with each balloon sounding, avoided dependence between successive
time blocks and permitted diagnosis in a physically meaningful way.

No significant difference was found between precipitation, gage by gage,
on seeded and unseeded days, even after deletion of 22 control days suspected
of contamination by previous seeding.

Of the original 715 three-hour blocks (365 treated, 350 control), 45
control cases were removed because of suspected contamination by previous
seeding under inversion and southeasterly surface flow conditions. Some
other blocks were removed where no clouds were present, or seeding had not
commenced, and some were lost in the pairing process. The final set con-
sisted of 216 six-hour blocks, 119 treated, and 97 control. The close asso-
ciation between the soundings and the six-hour blocks permitted detailed
stratification by physically meaningful cloud and air mass categories, a
valuable asset not available in the experimental day set where many types of
cloud and air masses could be experienced within one day.

J

:.

In addition to stratification by cloud/air-mass categories, the control
data were used to develop multiple regression formulae for predicting preci-
pitation from various sounding-derived cloud and air-mass parameters. These

-10-