Laserfiche WebLink
<br />B-11 <br /> <br />The observed. effect of seeding on rainfall appeared positive but far from significant. <br />* Colorado Rive:r * <br /> <br />Severe operational problems, and a need for post hoc categorization of days, leave the pos- <br />sible strength of evidence from this experiment in doubt. (After post hoc categorization the, <br />apparent direction of effect was consistent with other experiments carried out in the Colorado <br />Rocky Mountains. See Section 45 for a brief discussion of reanalyses of these earlier experi- <br />ments.) <br /> <br />* Israeli 11 * <br /> <br />Randomization tests showed the mean value of the double ratio statistic to be significantly <br />different from that expected under null conditions b~:yond the .05 level, allowing rej~:ction of <br /> <br />the null hypothesis in favor of the alternative hypothesis that rainfall was increased by seeding. <br />The observed average increase in rainfall under seeding was about 15 percent. <br /> <br />Some separations were made in subsidiary analyses. Target area separations showed rain- <br />fall increases under seeding in the interior sectors of the target to be greater than thOSI~ for the <br /> <br />entire target. Separations based upon cloud-top temperatures support the a priori expectation of <br />more reliable effects of seeding for temperatures in the range -lOoC to -150C than for tem- <br /> <br />peratures outside that range. <br /> <br />These results and conclusions are based upon a single summary report of the Israeli II <br />experiment (Gagin and Neumann, 1976). More detailed reports and the data on which the <br /> <br />reports are based are yet to be made available to the scientific community. If closer critical <br />analysis of the study and its data fails to detect flaws, the results of this study would be judged <br /> <br />to consitute confirmatory evidence that rainfall amounts have been increased by cloud s,eeding. <br /> <br />Published objections to the design of the earlier Israeli I, though natural in view of the <br /> <br />incompleteness of the early publications, do not now seem to us to be important, in. view of <br /> <br />later information. If the latter part of Israeli I can be taken as a confirmatory phase, as we <br />would suggest, it is possible that, after further critil:al study, Israeli I and Israeli II will be <br />judged to provide the first successful example of confirmation followed by reconfirmation. <br /> <br />* Alberta hail * <br /> <br />The disappearance of randomization leaves inadequate data for useful results or conclu- <br /> <br />sions. <br /> <br />* Santa Barbara][l * <br /> <br />The overall picture is still somewhat confused, with various analyses of phase I hovering <br />around the 5% significance level. The results are suggestive and en~ouraging, but positive con- <br />clusions do not seem to be justified. <br /> <br />* South Africa hail * <br /> <br />The results are suggestive of hail reduction but the results are not persuasive to us due to <br />the lack of rand.omization and to possibly important subjective biases, noted earlier. <br />Apparently, this program never was intended to be an experiment. <br />* North Dakota * <br /> <br />Overall diffefl~nces in rain for seeded and unseeded days are negligible and nonsignificant. <br />When seeded days are separated into days judged suitable and unsuitable for dynamic seeding, <br />the mean differenl~e in average rainfall between (suitable) seeded days and un seeded days <br />