Laserfiche WebLink
<br />A Statistical ~eanalY5is of the ;t.e?licated Cli::.iax I and II <br />Uintertime Orographic Cloud Seeding Experiments <br /> <br />Paul U. Mielke. Jr.l. Glenn ~_ Brier2. Le~is O. Grant2, <br /> <br />. 2,3 2 <br />Gerald J. Hulvey ,and Paul N. Rosenzweig <br /> <br />Colorado State University, Fort Collins. CO 80523 <br /> <br />Abstr.1ct <br /> <br />A reanalysis of the Climax I and II experiments is described. The <br /> <br />concern prompting this reanalysis is a suggestion arising from Colorado <br /> <br />State University analyses of extended area effects. Those analyses suggested <br /> <br />a region-~ide pattern of precipitation that, by chance. may have favored the <br /> <br />randomly selected seeding days for some of the important meteorological <br /> <br />partitions used in earlier analyses. In order to address this concern, this <br /> <br />reanalysis employs excellent covariate relationships developed before the <br /> <br />initiation of Climax II and ",hieh account for over half of the target <br /> <br />variability for most meteorological partitions of major interest (e.g., <br /> <br />warm SOO-mb temperatures and southwest 700-mb wind directions). The <br /> <br />statistical evidence of seeding induced increases associated ~ith this <br /> <br />reanalysis is generally much stronger than the previous analysis ",hich did <br /> <br />not utilize the covariate data. For example, the joint one-sided Wilcoxon <br /> <br />test statistic P-value for testing the null hypothesis that seeding did not <br /> <br />induce a precipitation increase during "'arm 500-mb temperatures of the <br /> <br />Climax I and II experi~ents is no'" 0.0013 compared to 0.0550 ~ith the <br /> <br />previous analysis \Jhfch did not utilize the covariate data. However, the <br /> <br />reanalysis also indicates tha.t previous estimates of increases attributed <br /> <br />to seeding based strictly on ratios of seeded to non-seeded precipitation <br /> <br />amount means are apparently too large. For example, the esti~'ted <br /> <br />precipitation increase of the combined Climax I and II data for the warm <br /> <br />500-mb temperature partition is reduced from 41 percent to 25 percent when <br /> <br />using the full set of J3ta as ~rnploY2d in this re~nal~sis. <br /> <br />-19- <br /> <br />A-1l-3 <br />