WSP13032 - Laserfiche WebLink

Home Browse Search

. e e - DRAFl' OF JUNE 8, 1992 - COLORADO RIVER BASIN EXPERIMENTAL SITE SELECTION 1. OVERVIEW OF EXPERIMENTAL DESIGN Several so-called "statistical" seeding experiments have been conducted with winter orographic (mountain-induced) clouds in the Western United States, particularly during the 1960's and 1970's. Optimism that cloud seeding may be able enhance the winter snowpack is largely based on statistical analyses of these experiments, and of some long- . term operational projects. However, several statistical experiments have been inconclusive, or have had their results challenged on physical grounds. Cause and effect relationships between seeding and increased precipitation have not been fully documented as pointed out in the recent AMS (American Meteorological Society) Policy Statement on weather modification (AMS, 1992). The Policy Statement recommends that, "The physical processes and specific conditions under which it is possible to increase, decrease, or relocate precipitation should be fully defmed." It has become clear that much improved "physical" experiments will be needed to validate and quantify seeding effectiveness. Past statistical experiments all had a physical component. Without physical observations, statistical analyses could not be applied. The problem has been the balance between physical observations and statistics. Most past cloud seeding experiments were of the "black box" type. While a detailed physical hypothesis may have been stated, actual observations were usually limited to precipitation at the ground with, perhaps, some supporting information on general cloud conditions (e.g., cloud top temperature, wind velocity). Thus, most physical processes took place unobserved in a "black box". If the statistical analyses did not indicate a seeding effect, there was no way to determine what went wrong. The physical hypothesis may have been incorrect, application of the design may have been faulty (e.g., frequent failure to target the seeding agent into the cloud), statistical testing may have been insensitive, the experiment may have had a "bad draw", or all of the above. A further problem with "black box" statistical designs is the difficulty of transferring the results to other locations. In recent years, improved instrumentation and knowledge have allowed much more emphasis on observations of the key processes involved in the physical chain of events following seeding. These include monitored release of the seeding agent, transport and dispersion of the agent, nucleation of ice crystals within the SL W (supercooled liquid water) cloud, subsequent ice particle growth by diffusion, accretion and aggregation, and fallout of snowflakes to the surface. A limited number of what we will call "direct detection" experiments have been carried out since 1985, which have validated the winter orographic cloud seeding hypothesis in particular conditions. These types of experiments need to be expanded in terms of better design and instrumentation, and in terms of application to a wide range of meteorological conditions. A much improved statistical/physical experiment can then be designed and conducted to quantify the seasonal increases to be expected from properly conducted seeding. 1 d,-k- /Uu.5J L /"" c4'<.A-~ -Fl>r c>" ~~~7::~) ~ r 0- I d~ k./j 10 7~>0- ~ d,re.cj'! dec. ';"/~,/ f)~crve- 9ooi) {-la, lirA.. wtl~,,1