Laserfiche WebLink
<br />Table 3.2. Double ratio results for the northwest control, the west control, and the <br />"statistical" control. <br /> <br />Northwest <br />Control <br /> <br />West <br />Control <br /> <br />Statistical <br />Control <br /> <br />--------------------------------------------------------- <br /> <br />Mav-AuJUSt <br />Target Seed/Control Seed <br />Target No Seed/Control No Seed <br />Double Ratio <br /> <br />.92 <br />1.09 <br />.84 <br /> <br />1.22 <br />1.38 <br />.89 <br /> <br />.96 <br />1.12 <br />.86 <br /> <br />was less in the seed years, but the deficiency in the target-area was more pronounced. As a result, <br />the double ratios for the May-August season are less than 0.90, which suggests that rainfall in the <br />target area during the seed years was less than would have been expected naturally. No attempt has <br />been made to establish the statistical significance of these results. <br /> <br />It is important to note that the probabilities presented in this report are accurate only when the seed <br />and no-seed cases are chosen at random. When the seed cases are from a single seed period and <br />the no-seed cases are from an earlier historical period, as is the case here, climatic trends can lead <br />to differences between seed and no-seed cases that have nothing to do with the actual seeding. <br />Gabriel and Petrondas [8] demonstrated this effect, using actual rainfall data, in a number of <br />experiments simulating cloud seeding projects. Therefore, the comparisons tend to overestimate <br />the probability that apparent seeding effects are real. <br /> <br />3.4 Large-Area Study <br /> <br />The gridded rainfall field provided a rainfall estimate for every grid cell for every month <br />(April-September) for every year (1948-85). Seed/no-seed ratios were computed for each of the <br />704 grid cells, using the mean rainfall for the seed period (1975-85) and the mean rainfall for the <br />no-seed period (1948-74). Ratios were computed for the May-August "season" and for each <br />month, April through September. The Wilcoxon rank-sum test (Dixon and Massey [9]) was used <br />to determine the significance of each of the seed/no-seed ratios. <br /> <br />3.4.1 May-August "Season" <br /> <br />The vast majority of the seeding took place in the 4-month period of May through August, <br />Seed/no-seed rainfall ratios for the May-August season are shown in figure 3.6. Only four grid <br />cells have seed/no-seed ratios greater than 1.1, all to the west or north of the target area. A <br />crescent-shaped band from the southwest corner through the center of the study area to the <br />southeast comer contains ratios between 0,75 and 0.9. Almost all of the target area is included <br />within this band. A few grid cells with ratios between 0.5 and 0.75 are found in the southwest <br />quadrant of the target area, and four grid cells with ratios between 0.9 and 1.1 are found in the <br />southeast comer of the target area. <br /> <br />The rank test shows that the seed/no-seed ratios (with a few exceptions) in the western 60 percent <br /> <br />19 <br />