WMOD00400

• Contract/Permit #: 14-06-D-6801
• Title: Structure and Seedability of San Juan Storms
• Date: 11/1/1976
• Weather Modification - Doc Type: Report

I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I

56

seeding th~ dissipation stage.
2. Discussion of the statistical experiment

This picture predicts a positive effect for unstable stages and no effect

for stable stages, so the net effect should be a positive seeding effect.

This \vas not observed in the San Juan proj ec t. However, not all storm stages

\vere seeded. Instead, only those storm periods were seeded during Hhich it

Has forecast that the 500 mb temperature Hou1d be -23C or warmer. (See

Hjermstad, 1975, for a more exact specification of the seeding criteria,

\oJhich varied some\oJhat over the five year period.) Since the early stable

and neutral stages of the storm had Harm 500 mb temperatures relative to

the later unstable stage, the requirement that the 500 mb temperature be

warm tended to favor the stable and neutral stages over the unstable stages.

HO\oJever, as the 500 mb temperature cooled the cloud top temperature warmed,

so that the cases Hith the \oJarmest cloud top temperatures were excluded.

More importantly, these criteria tended to exclude the more seedable unstable

stages and to include the unseedable stable stages in the statistical experiment.

The statistical experiment was diluted by including large numbers of

stable, unseedable storm stages, and also by the phenomenon of carryover

seeding. We observed (as did Hobbs et a1. 1975), that the AgI often remained

in the experimental area for 12 hours or more after the generators Here

turned off, because of the pooling of the AgI ahead of the mountain range

under blocking flow conditions. This led to the release of AgI into storms

considerably after the seeded period, and nominally unseeded storms were in

fact often seeded.

The effects cited above should lead to a dilution of the statistical

experiment, but not to a null effect. The arguments presented here are not

meant to imply that the seeding technique should not have worked, but only

that the statistical experiment wis seriously weakened by the inclusion of