OCR Text |
Show used ext_ens~vely _during the 1968-69 operational season. This model was used in combmation with an atmospheric diffusion model to predict the time of arriv~l of seeding material at various target area locations, both for ground and airb~rne 1\gI rele~s:s .. Time histories of supporting physical measureme~ts (silver-in-pr~c i~itahon and the number of snow crystal replicas possessi~g AgI cent_ers) mdicat_e that the model show£ quite good skill in predicting the time of arrival of seedmg material into specified areas. Thus, the orographic flow model in conjunction with the diffusion model proved to be a valu able analysis aid. However, when the predicted orographic precipitation distribution was corn pared to observations over the project network, very little correlation was found. This was partly due to an inadequate model approximation of the actual Park Range ridge profile. Thus, additional calibration of the precipitation portion of the mode 1 should be made by varying the mountain profile and also the crystal growth-fall time curves. 6. 5 Effectiveness of Experimental Designs and Evaluation Methods Two main experimental designs were employed in the Park Range Project. These were (1) pulsed seeding (1 cph) from Emerald Mountain, and (2) randomized seeding of either the first or fourth hour of a designated six hour operational block, using either Mt. Harris ground generators or airborne pyrotechnics. The pulsed seeding design was used routinely for the first two years of seeding operations, and with decreased pulsing frequency (1 / 4 to 1 / 2 cph), during most of the third. Power spectral analysis techniques were invoked to identify pulsed seeding effects in target area precipitation rates. A variety of complicating factors, especially the complexity of low le vel diffusion processes, prevented conclusive results under this design. On several occasions when sufficiently steady wind and natural precipitation regimes prevailed, the pulsed character of seeding effect was identifiable. There are also indications, both from model trajectory calculations and from seasonal snowpack on two downwind ridges, that much of the seeding effect from releases at Emerald Mountain was felt downwind of the main target area. The randomized seeding design used in the 1968-69 season allowed the collection of both seeded and unseeded data within each six-hour operational block. Two main types of analysis routines were used ·o n 1968-69 data samples. The first of these routines included stratification of collected data into seeded and unseeded samples (by using trajectory calculations for seeded crystals), and statistical testing for significant differences in sample meq.11s~ In the ~ase of snow rate samples, additional stratification was made accordmg to estl~~ted cloud top temperature categories. Using these stratifications large positl ve 333 |