Laserfiche WebLink
<br />002537 <br /> <br />previous project of this type has amassed as much data of such high <br />quality." <br /> <br />The difficulty of analysis and evaluation was increased because post <br />facto data analysis showed many of the clouds actually seeded were <br />not of a suitable type. In addition, on some days weather did not <br />develop as forecast. In numerous cases, the seeding agent remained <br />in the area beyond the planned experimental seeding periods. On <br />other days rapid weather changes produced conditions under which <br />seeding decreased precipitation. <br /> <br />The initial statistical analysis found no difference between precipi- <br />tation on seeded experimental days and control days. However, when <br />days of missed forecasts were removed, and data from experimental <br />days were reduced to 6-hour time blocks rather than 24-hour periods <br />to improve the correlation between meteorological conditions and <br />precipitation, increases during certain classes of seeded cases were <br />statistically significant. <br /> <br />The evaluation was based on the physical factors influencing seeding <br />effectiveness, such as the supply of cloud water, the supply of <br />natural nuclei, the trajectory of falling snow, and the degree of <br />thermodynamic instability of the clouds. This enabled evaluators to <br />examine the role played by each factor and seek a method for trans- <br />ferring this knowledge to mountain barriers having different cloud <br />climatologies and different terrain features. <br /> <br />. , <br /> <br />The evaluation took two forms: (1) the initial formal statistical <br />analysis based on the precipitation data for 71 experimental treated <br />days and 76 experimental control days and (2) a statistical analysis <br />of precipitation and upwind sounding data arranged in 6-hour time <br />blocks. <br /> <br />Constraints imposed by randomization and by suspension permitted the <br />two primary goals of the project to be achieved with differing <br />degrees of success. The post hoc analysis demonstrated that winter- <br />time precipitation had been increased by seeding under certain cloud <br />conditions and decreased under others. Specifically, when cloud tops <br />were warmer than -29 .C, seeding apparently resulted in a decrease in <br />precipitation; but this has not been statistically verified, and with <br />light winds there were increases in this cold category. Orographic <br />clouds containing deep embedded cumulus clouds showed little differ- <br />ence between seeded and unseeded samples. <br /> <br />Investigators concluded the full design potential of the CRBPP was <br />not realized because many clouds actually seeded were not of a <br />suitable type. To a considerable extent, this occurred due to the <br /> <br />11-2 <br />