Laserfiche WebLink
<br />VI. STATISTICAL HYPOTHESES <br /> <br />The statistical hypotheses to be tested in the HIPLEX-1 experiment <br />are derived from the conceptual models (section V). The basic null <br />hypotheses to be tested are: <br /> <br />HI. <br />o . <br /> <br />Seeding does not affect the probability of rain at <br />cloud base. <br /> <br />Seeding does not affect the amount of precipitation <br />at cloud base averaged per experimental unit. . <br /> <br />The statistical hypotheses relating to the steps in the precipitation <br />process involve 13 primary response variables representing 10 differ- <br />ent physical characteristics. Eight secondary response variables <br />not considered essential in validating the conceptual models are also <br />included. Five final test statistics complete the a priori statisti- <br />cal hypotheses. The sixth element of the conceptual models, dealing <br />with rain at the ground, is not included in the statistical hypotheses. <br />The small cumulus congest us clouds produce only small amounts of <br />precipitation over small areas, and it is impractical to install a <br />gage network dense enough to estimate adequately the rain from <br />individual clouds. <br /> <br />H II. <br />o . <br /> <br />The elements of the statistical hypotheses are defined in table 3. <br />The individual null hypotheses and the expected alternative hypotheses <br />representing the steps in the physical process are presented in table 4. <br /> <br />More than one approach to estimating both the time to initial precip- <br />itation and the rainfall is necessary because no one instrument or <br />procedure is completely satisfactory at this time. Section VIII <br />outlines how the variables are measured. <br /> <br />.. <br />.; <br /> <br />10 <br />