WMOD00464

• Title: The Management of Weather Resources - Volume II
• Prepared For: The Weather Modification Advisory Board
• Date: 6/30/1978
• Weather Modification - Doc Type: Report

C-2

-------.

'--

As Fisher once wrote, "5 per cent. point...a scientific fact should be regarded as experi-.
mentally established only if a properly designed experiment rarely fails to give this level of
significance" (page 504 of Fisher, 1926). If a second confirmatory phase is to serve its pur-
pose, it must be conducted for enough years to provide adequate precision, just as if it were a
first confirmatory phase.

In any case, the existence of the first confirmatory phase, conducted consistently over the
appropriate number of years, could be regarded as fixing the confirmatory part of the new phase
so solidly as to make embedding it in a parallel exploration reasonable.

* piggyback experimentation *

Today the basic questions are unresolved after major efforts conducted with increasing
care. We have just emphasized the need for even more care in experimental. weather
modification. How then can we gain worthwhile information from experimentation piggybacked
on operational cloud-seeding if this is feasible at all?

It is our firm belief that piggyback experimentation cannot be cost-effective unless it is
conducted with the greatest care. Specifically, each and everyone of the following are essential:

1) randomization

2) blindness of decision-makers and data collectors (a suitable version of OJR-blocking -. see
Sections 15 and 20 -- may suffice for the decision-makers)

3) auditing by externally employed and supervised personnel

4) firm advance agreement as to what measurements, in what combination and how
analyzed, give the main answer

5) at least one analysis in terms of re-randomization, preferably by two-stage randomization

6) availability of fully detailed data to those who may wish to reanalyze it, from either
confirmatory or exploratory points of view.

We are doubtful of the acceptibility of such a package by those concerned in a weather-
modification operation, but we would not regard anything less as a worthwhile contribution to
EXPERIMENTAL KNOWLEDGE.

* should we go forward? *

Anyone with a fervent conviction, EITHER that cloud-seeding can be useful in enhancing
rain, OR that it cannot, can so interpret the available data to find it supporting his or her posi-
tion. This has been true for at least two decades, over which period the plausible extent of the
increase in rainfall has decreased somewhat.

We suggest, in view of the great importance of enhancement if it exists, that it is quite
reasonable to go forward with experimental weather modification by cloud-seeding, but only if
(i) the experiments are conducted with the greatest care and (iO it is generally understood that
the decade, not the year, is the time step within which we hope to make progress in determin-
ing whether cloud-seeding is useful in enhancing rain.

* world-wide steps *

At a time when world-de planning for weather modification experimentation is taking
place, -we suggest that it is important to consider what might be done to confirm or re-confirm
the most promising and definitive rainfall enhancements so far obtained. A direct repetition of
Israeli I seems unlikely, now that the experiment has been converted into an operation. (To a
degree, Israeli II was such a repetition.) We have no idea of the feasibility of a repetition of the