WMOD00385 - Laserfiche WebLink

Home Browse Search

be random. The difference in the amount of precipitation from seeded and from unseeded clouds is' the effect of seeding. Once again, vari- ables other than precipitation can be used. Either the target-control or the seed-no seed design can give s~Ltis- factory results wi thin certain limitations. The assumption that: the target and the control experience the. same climate is not always valid, and the correlation between them is not necessarily constant froll stOrM to storm or from year to year. There is a chance th~Lt the randoll selection of no-seed days will result in more naturally heavy precipitation days in one category or the other, thus giving distorted results. Furthermore, the seed-no seed design automatically reduces -the ..unt of precipitation augmentation that can be expected in the target area. A variation of these designs is the comparison of precipitation from seeded clouds with long-term precipitation av~rages for the sam~~ area. In this case all seedable storms are seeded, but the asst~p- tion that precipitation during the seeded years would have been com- parable to the long-term average had seeding not occurred is questionable. The purpose of most orographic seeding projects is to increase the amount of water stored in reservoirs or to increase the flow of a river. A certain amount of the precipitation will evaporate before melting and running off into the streams. Some will go :into th~~ soil and be evaporated, transpired, or stored. However, the scarcity of data available each season makes hydrologic evaluation more difficult. The choice of evaluation techniques is based on a thorough study of the area's climatology. If the correlation between two areas is consistently high, then target-control may be acceptable. If a large number of seedable storms can be expected each year, then seed-no seed may be preferred. If there is little year-to-year variation in precipitation, then the long-term average may be ml acceptable control. In general, however, one finds insufficient climatological data to make a clear choice. In this case, most statisticians suggest the seed-no seed procedure. ;, The amount of data needed to determine the results of seeding Wil th a high degree of confidence depends on the natural variability of the data, their accuracy, how representative they are of the whole area, and the magnitude of the effect to be detected. Increases in wilnter season precipitation on the order of 5 to 30 percent are to be expected from today's technology. As a general rule, about 100 seed- able events will provide enough data to reliably detect an incroase 5