Six Hour Analyses of the Bridger Range Randomized Winter Orographic Cloud Seeding Experiment

Abstract

The Bridger Range winter orographic cloud seeding experiment was conducted during the early 1970s. Published post hoc exploratory statistical analyses used 24 h experimental units. However, 6 h precipitation observations exist which have not been previously tested with non-parametric statistics. They should be better represented by available 6 h partitioning data. This experiment produced high quality precipitation data and was one of few withassociated physical studies adding credibility to statistical suggestions. Use of control gauge data substantially reduces natural variance in target precipitation. Two independent statistical approaches were applied to the 6 h dataset. Results strongly suggest that seeding was effective when conditions were conducive to orographic cloud formation with nearcrestline temperatures sufficiently cold for adequate nucleation with silver iodide. Specifically, the null hypothesis (seeding had no effect) was rejected with one-tailed P-values near 0.001 for the single partition of seeded zone temperatures less than the median. That subpopulation was further reduced by about 50% with the requirement of rawinsonde observations, launched only when clouds existed near or below crestline elevations. Similar very low P-values resulted from this dual partition with much reduced sample size, and for an even smaller population with700 mb dew point depressions less than their median value. These results are physically reasonable.Partitioning by cloud-top temperature and cloud thickness suggested that seeding could be effective even when thick clouds with cold tops were present.  It is stressed that these results are based on post hoc exploratory analyses so they can only be viewed as suggestive and not conclusive proof. Suggestions are given for future randomized winter orographic experimentation.