Laserfiche WebLink
<br />3.3 Target-Control Comparisons <br /> <br />The Western Kansas Weather Modification Project was designed to seed each day that seedahle <br />clouds occurred in the project area, and all seedable clouds that could be reached with the available <br />aircraft were treated. Our analysis of the effects of seeding on rainfall at the actual obselVation <br />sites is based on relationships between target-area rainfall and control-area rainfall and <br />comparisons between the seed years and the no-seed years. <br /> <br />The first step in the analysis was to derme one or more control areas (that is, unseeded areas <br />whose rainfall is closely related to the rainfall in the seeded target area). Because the majority of <br />storms affecting the target area move from a generally southwest or northwest direction, only <br />observation sites west of 101 oW were considered for control stations. Of these sites, only those <br />with fewer than nine missing years out of the 1948-74 no-seed period were used. Mean seasonal <br />rainfall for each of 92 candidate stations was correlated individually with a mean seasonal <br />target-area rainfall, defined as the mean rainfall of 14 stations located within the target area (again <br />restricted to those with fewer than nine missing years out of the 27 no-seed years). Appendix 1 <br />lists the 14 stations used for target-area precipitation calculations. <br /> <br />Two kinds of controls were selected. The rll'St included all stations within two geographical areas <br />that included the clusters of stations showing maximum correlation with the target rainfall <br />(typically values of 0.7 or greater). Two such areas were identified, one to the northwest of the <br />target area and another to the west (see figure 3.1). The second kind of control was dermed as <br />that group of 16 stations (west of 10l0W) with correlation coefficients greater dfan or equal to <br />0.67. The station names and correlation coefficients are listed in appendix 2. The mean rainfall <br />values for the target and the three controls are presented in table 3.1. <br /> <br />Table 3.1. Mean seasonal (May-August) rainfall (inches) for target and <br />control areas. . <br /> <br />Seed <br />1975-85 <br /> <br />No-Seed <br />1948-74 <br /> <br />Difference <br />S-NS <br /> <br />Percent <br />(of NS) <br /> <br />------------------------------------------------------------- <br /> <br />Target <br />NW Control <br />W Control <br />Statistical Control <br /> <br />9.5 <br />10.4 <br />7,8 <br />9.9 <br /> <br />11.6 <br />10.6 <br />8.4 <br />10.3 <br /> <br />-2.1 <br />-0.2 <br />-0.6 <br />-0.4 <br /> <br />-18 <br />-2 <br />-7 <br />.- 4 <br /> <br />The double ratio <br /> <br />Tarset-seedlControl-seed <br />Target-no seed/Control-no seed <br /> <br />was used to evaluate the effects of seeding on observed rainfall in the study area. The douhle ratio <br />should minimize.the effect of climatological variations so that a naturally wetter no-seed period <br />would not indicate that seeding had a negative effect on rainfall. It assumes that the same <br />relationships between target and control areas exist in both seed and no-seed periods. <br /> <br />Results are presented in table 3.2. Rainfall in the target area and also in the three control areas <br /> <br />18 <br />