Laserfiche WebLink
<br />inconsistencies and dcficiencies in thc statistical and physical evidence obtained thus far. <br />Progress in physical understanding comes from noting the unexpected and following it up as well <br />as from confinning the expt:cted. Mindful that the results from Cl posfer;ori analyses might <br />evince a physically interesting result that in fact might only rellect chance, strong statistical <br />support for a result alerts the physical scientist even though there is no ready theory to explain <br />the results or the findings run counter to the postulatcd seeding conceptual model or the findings <br />appear to be inconsistent with thc findings of previous physical studies. Physical undcrstanding <br />is clarified and advanced through follow-up statistical and physical studies and experiments <br />prompted by such findings. <br /> <br />Cl, SfafiSfical,\,tlldie,,' <br />First and foremost, it is esscntial that statistical studies be conducted to conlinn or revise, <br />as necessary, the statistical results of this preliminary study. The results of this baselinc study <br />will idcntify when significant changes in seeding effectiveness occurred and the nature of those <br />changes, Thcse results could servc as a starting point for the studics to determinc the physical <br />causc(s) of these changes in sccding etTectiveness in particular and for thc cfforts to optimize the <br />cloud seeding elTorts in general, The following conlirmatory statistical studies arc needed: <br />I). Pcrhaps the biggcst conccrn in the preliminary study is that all the operational seeding <br />programs were evaluated using the same control site, the Merced River at l)ohollO Bridge <br />(MlW). Although MDP had a relatively high correlation with all the target sites, evaluations <br />using controls that are more geographically rcpresentative of their respcctive targcts and have <br />higher statistical corrclations with their rcspective targets would provide morc precise results, In <br />addition. the evaluations with othcr controls would provide a means of contimling (he <br />preliminary study results. Toward these ends, the most representative controls with a long record <br />of FNF data (hat arc available for each of the opcrational seeding program targets should be <br />idcntified and used in are-evaluation of each of the operational seeding programs, <br />2,) Evaluations of additional targets in each of the operational seeding programs would <br />pro\'ide annther means of confirming the preliminary study results, If those targets werc better <br />located with rcspect to thc arca being alTccted by the sceding, a morc precise cstimate of the <br />magnitude of the sccding efTect would bc obtained, Evaluations of other possible targets for a <br />given operational seeding program would also provide a first estimate of thc arca distribution of <br />the sceding effect for that operational seeding program, Toward these cnds, additional targets <br />with a long record of FNF data for each of the operational sceding programs should be identified <br />and used in supplemelltary evaluations of the operational seeding programs, <br />3), The analysis of the time evolution of the secding signal to detennine \....hen significant <br />changcs in seeding effectiveness occurred should be carried out for all the evaluations.. As in the <br />preliminary study, the progressivc statistical cvaluation using ratio statistics, called the <br />cumulative year statistical evaluation method. \....ill be used to revcal the trcnds in seeding <br />elTectiveness as a function of time. A signilicant change in trend in the plot of the cumulative <br />year evaluation results as a function of the cumulative numbcr of operational years should be <br />indicative of a significant change in some aspect of the mcteorology and/or seeding procedure <br />that alTected seeding efTectiveness. To better reveal these trends. a moving average statistical <br />evaluation method using ratio statistics will also be applied to the evaluations. According to the <br />moving average statistical evaluation method, the seeding is evaluated in a sclected block of <br />years that moves through the period of operations (for example, a 10-ycar block of years starting <br />with years I thm 10, then years 2 thru 11. then years.3 thru 12. etc). <br /> <br />48 <br />