Laserfiche WebLink
<br />3. Steps 1 and 2 will be repeated for all of the permutations that are possible after <br />following the protocol described above, and a sorted vector of the SR's obtained in the <br />permutations will be produced. <br /> <br />4. The fraction (P, in percent) of the randomly obtained SR's which are larger than the <br />observed one in the real experimental allocation is the chance that the experimental <br />result is caused by the natural variability. This chance is referred to as the <br />"significance leveL" <br /> <br />2.11.5 Seed versus no seed comparisons <br /> <br />Several analyses will be pursued in the S versus NS cell comparisons. The first will be <br />comparison of means, SR, and DR, if appropriate, for the various cell properties as a function <br />of treatment decision. This analysis will be done for the "long track" calculations, and the <br />results will be stratified by the number of flares and time of treatment. In addition, plots of <br />mean cell properties versus time for the Sand NS samples will be constructed. Finally, time- <br />height reflectivity plots will be composited for the Sand NS cases. The time of the <br />qualification pass will be used as the "0" time reference in all time plots. All analyses will <br />be done for both approaches. <br /> <br />Once the cell comparisons have been completed, analyses for seeding effects beyond the scale <br />of individual cells will be made. The first will look at the merger of cells to see whether <br />merger is more prevalent in S cells than in NS cells. In Texas, the S cells merged more often <br />than the NS cells (Rosenfeld and Woodley, 1989; 1993) in agreement with the conceptual <br />model. The same result is expected in Thailand. <br /> <br />The next analysis will be the "focus,ed area" approach that involves calculation of cell <br />properties for various radii around each treatment position as a function of time. In Texas, <br />the effect of seeding appears to begin on the scale of the individual cell and spread outward <br />with time (Rosenfeld and Woodley, 1989; 1993). Using this "focused area" approach, which <br />is discussed extensively by Rosenfeld and Woodley (1993), it will be possible to determine <br />whether this effect exists in Thailand. <br /> <br />Finally, the experimental units containing all of the cells, including those that did not receive <br />either AgIoI' simulated AgI treatment, will be examined for seeding effects. Each <br />experimental unit covers about 2,000 square kilometers, and documenting seeding effects on <br />this scale will be more difficult, as was shown in the section dealing with the duration of the <br />experiments. <br /> <br />21 <br />