Abstract
A randomized exploratory single-area cloud seeding experiment was carried out in the Bridger Range of southwestern Montana during the winters of 1969–72. Seeding was accomplished using ground-based silver iodide (AgI) generators located more than midway up the west (windward) slope of the north–south Main Ridge, thereby avoiding trapping by lower stable layers. A secondary ridge from 5 to 20 km east of the Main Ridge was the expected target. An extensive airborne plume tracing program provided strong evidence of successful targeting of the AgI seeding material, with further evidence furnished by tracking of pibals and silver-in-snow analysis.
The experimental unit was 24 h beginning at local noon, a natural diurnal minimum in precipitation intensity. The response variable was daily precipitation amount as measured by a dense network of recording gages. Locally-launched rawinsondes and a thermograph atop the Main Ridge provided data for partitioning the experimental days.
A post hoc statistical analysis was conducted utilizing upwind and crosswind control gage data. Results from both the Wilcoxon rank-sum test and the recently developed multiresponse permutation procedure (MRPP) strongly suggest that increased target area snowfall resulted from seeding when AgI plume temperatures were colder than approximately − 9°C. Double ratios yielded estimates of ∼ 15% more seasonal target area precipitation than predicted by control gages on nonseeded days, while a target-control analysis of independent snow-course data strongly suggested seeding enhanced the seasonal snowpack by more than 15%.
Consideration of plume tracing findings and AgI generator calibration results suggest that the artificial ice nuclei concentration in the seeded volume would be quite limited at temperatures warmer than approximately − 9°C. This provides a plausible physical explanation for the results suggested by the statistical investigations.
