Laserfiche WebLink
<br />e <br /> <br />V. STATISTICAL HYPOTHESES <br /> <br />. <br /> <br />The statistical hypotheses to be examined in the SCPP-1 experiment are <br />derived from the conceptual model. The hypotheses relate to the steps in the <br />initiation of the precipitation process and involve seven primary response <br />variables. Three response variables were selected to document the initiation <br />of the ice phase and subsequent crystal growth mechanisms in the cloud to <br />precipitation size particles. Thi s wi 11 be determined by aircraft. The <br />second two response variables will document the rapid initiation and subse- <br />quent fallout of precipitation due to treatment. These will be determined by <br />a 5 cm radar. Finally the last two response variables will identify changes <br />in precipitation as measured at the surface by gages. <br /> <br />Both the aircraft and radar response variables have been selected as being <br />the most sensitive to treatment effects, as well as being reliable measure- <br />ments. Other response variables such as using oil-coated slides to determine <br />crystal habit, aggregation, riming, etc., have been suggested but do not have <br />the necessary reliability. They have thus been reduced to secondary response <br />variables. Another philosophy followed in selection of response variables <br />was to select a variable from one measurement platform that would correlate <br />highly with the next measurement platform in the series. Thus the so-called <br />IIhandoffll vari ab les incl ude both the 2DP ice crystal concentrations and <br />measure of the ice water content (IWC) of the crystals by aircraft and the <br />radar determined area of the 20 dBZ contour 30 minutes after treatment. It <br />is envisioned that if these handoff variables show little or no response that <br />the subsequent measurement platform would be unable to determine a seeding <br />signature. <br /> <br />The response variables used to test the hypotheses are defined in table 5-1. <br />The null hypotheses for each of the response vari ables and the expected <br />alternatives are presented in table 5-2. <br /> <br />. <br /> <br />The aircraft response variables will be measured relative to individual <br />cells. Since data from two or more cells normally will be obtained, the test <br />statistics will use the mean value of the response variables for all cells <br />measured in each experimental unit. For purposes of radar evaluation, two <br />floating boxes will be identified, centered at 15 and 30 minutes downwind <br />of the seeding line. The seeding effect is assumed to be advected IIdown- <br />wi nd II at a rate based upon cell mot ions as observed by radar. The box for <br />the vari ab le R15A15 extends from 7 to 22 minutes downwi nd, centered on the <br />15 minute projection of the line laid down at t. Thus it would include <br />only one seeding line. This also allows the R15A~5 response variable to be <br />sensitive to the earlier initiation of precipitation (i.e., time to first <br />echo) in the treated cases. The box centered at to + 30 minutes is eval- <br />uated at 30 minute intervals and contains three seeding lines. Assuming a <br />new seed line begins every 10 minutes. The mean value of a given response <br />variable over the corresponding number of boxes will be used to characterize <br />the experimental unit. Detail s of the radar primary and secondary response <br />variables are given in updated Appendix B. <br /> <br />Precipitation response variables will be calculated from precipitation <br />observations within the EPS (evaluation precipitation swath). The EPS is <br />defined by determining a mean seeding line location and orientation. Assum- <br />ing that seeding effects are advected at a constant rate, which is determined <br /> <br />15 <br />