WMOD00464 - Laserfiche WebLink

Home Browse Search

B-6 ,,~ a typical experiment must be conducted over a period of several years. The longer the period of experimentation, the more difficult it is to maintain a standard set of experimental pro- cedures. The personnel may change, the operational methods may be altered as new technol- ogy becomes available, administrative (or political) constraints may vary, etc. Generally,. changes will tend to make the experiment less sensitive. Changes are often unavoidable, and even desirable during an exploratory phase. They are, of course, almost inexcusable in a confirmatory phase. · FACE · In the Florida area cumulus experiment, numerous changes were made in operations. Throughout, the scientist was airborne, in the plane responsible for delivery of the treatment. From 1970-1975, AgI flares were ejected on experimental days while no flares were ejected on control days; in 1976, either AgI flares or placebo flares were ejected ("because the flare ejec- tion could be detected on tp.e light aircraft"). Types of flares employed varied from time to time during the course of the study. The basic source of rainfall estimates also varied: in 1970 and 1971, radar from the University of Miami was used with manual generation of rainfall esti- mates, without adjustment in 1970, with adjustment in 1971. In 1973, 1975 and 1976 the' NOAA radar of the National Hurricane Center was used with digitized estimates of rainfall, adjusted daily by data from rain gauges. Unconstrained randomization was employed during 1970-1973 to assign experimental days to seedl:ld or non-seeded treatment conditons. In 1975, treatments were randomized within blocks of variable length but no greater than 5 (3 seed days and 2 control days). In 1976, res- tricted randomization was used, of a character unknown to the decision makers. The authors of FACE reports properly recognize that "because it has been an exploratory effort with evolutionary design, FACE cannot be regarded as a conclusive experiment." For a clear-cut confirmatory experiment, which FACE was NOT intended to be in 1970-1976, experi- mental conditions could not be allbwed to vary so greatly. · NHRE · Other than an attempt to mimic seeding operations carried out and extolled in the U.S.S.R., no physical basis actually existed for the national hail research experiment. Serious efforts were made by the NHRE staff to remedy this situation while the experiment was already in progress. (Although a great many concepts may have been originated and some insights may have been gained during this effort, there still appears to be no satisfactory physical basis for suppressing hail damage due to severe storms on the High Plains,) * Colorado River Basin * Operational problems were many, and severe. Methodological instructions from the Bureau of Reclamation varied from year to year, and sometimes within years: Of the 20 largest storms occurring during the first three winter seasons of the study, 14 wete excluded due to I operational restraints, 5 were unseeded and 1 was seeded. (If, as later exploratory analyses sug- gest, large storms, which are likely to have thick cloud layers, are unfavorable for seeding any- way, this point may not be of major significance,) * Israeli II * No exposition of operational details of the second Israeli cloud seeding experiment bey~d the brief paper of Gagin and Neumann (1976) has been made available to the Task Force. We know nothing to suggest that conditions were altered over the course of the