Abstract
Trace chemical measurements of the silver content of snow have been used to investigate the transport and dispersion of silver iodide cloud seeding aerosols into and around two large target areas in the central Sierra Nevada between 1978 and 1992. The background concentration of silver in snow samples in this region is extremely low [BAg = 2.0 parts per trillion (ppt); standard deviation, σ = 1.0 ppt], and the silver from the seeding activities is readily detectable. The studies, in winter snowstorm conditions, show that targeting of the seeding aerosols was modest to poor with large variability both spatially and temporally. Analysis of several thousand snow samples over a period of several years has demonstrated that only 20% (average) of the precipitation that fell within the intended target area during seeding activities contained silver from the silver iodide seeding aerosol above the assigned “threshold” concentration of BAg + 2σ = 4.0 ppt. Targeting of one of the catchment areas under southerly flow storm conditions was particularly poor (less than 8% of the sampled snow containing detectable silver above the “threshold” value).
Evidence is also presented of the transport of silver iodide in directions and into areas other than those intended including upwind control areas used for estimating seeding effects in the target. In one period of the study between 1987 and 1990, emphasis was placed on southerly flow storms. It was found that contamination of two control areas in the Lake Almanor region generally occurred in the early phases of these southerly flow storm periods when winds at the lower levels were from northeast to southeast prior to frontal passage. The method used for estimating the layer-averaged wind prior to storm classification for seeding purposes in this project placed equal emphasis on winds at all levels from 2000 m, the approximate elevation of the ground generators, to the −10°C level. Hence, aerosols released at ground level were often transported inadvertently toward the control areas by these low-level winds. The problems that these results raise in regard to the traditional use of precipitation statistics for assessing seeding effects are discussed. The results of the studies indicate that it would be necessary to produce very substantial precipitation changes in the limited areas where the seeding silver is present in the snowfall to yield a statistically acceptable change over an entire target area. The trace chemistry results presented may explain why one of the randomized cloud seeding experiments did not achieve statistically significant seeding effects and why another statistically significant result may have been misinterpreted.
New seeding program designs and assessment methodologies need to be developed that not only will produce better targeting, but also positively identify the precipitation that has been impacted by the seeding process, so that seeding effects may be quantified with a higher degree of confidence.
