Laserfiche WebLink
<br />characteristics at the surface. However, it should be supplemented by photographic documentation which <br />provides more detailed data, though requiring considerable manual reduction. <br /> <br />The degree of riming was shown to have been reduced in some of the reviewed experiments. This should <br />be a consequence of seeding if enough ice crystals are created to utilize most of the excess SL W. The <br />degree of riming usually is not discernible from aspirated imaging probe data so manual or photographic <br />observations are required at the surface target. <br /> <br />It is useful to obtain snow samples for silver analysis at frequent intervals to help evaluate seeding <br />effectiveness. Enhanced silver levels in the snow do not prove any seeding effect directly since most or <br />all of the silver could result from scavenging by natural snowflakes. However, fmding only background <br />silver concentrations very likely means the target was not impacted by seeding. Thus, the silver-in-snow <br />data provide a partial check on claiming real seeding effects that are actually natural variations. <br /> <br />Highly sensitive precipitation gauges are needed for physical experiments because the seeding effects may <br />be very brief (e.g" only a fraction of an hour for a single airborne line), and rates may be low with typical <br />SL W amounts. For example, the total precipitation amounts from individual seed1ines of AgI, reported <br />by Super and Boo (1988b), ranged from 0.10 to 0.22 mm. Conventional weighing gauges have a <br />resolution of 0.25 nun (0.01 in), and are unsuitable for physical experiments unless modified. <br /> <br />Radar has sometimes been used in attempts to follow the effects of seeding between lowest aircraft levels <br />and surface instruments. However, radar evidence of winter orographic seeding effects is nonnally <br />inconclusive unless the natural IPC is very low. That is because the radar reflectivity factor is directly <br />proportional to particle concentration, but is proportional to the sixth power of particle size. Thus, the <br />returned signal from a few large natural snowflakes can completely mask that pf an order of magnitude <br />increase in smaller seeded crystals. Seeding could conceivably decrease the radar returned signal while <br />increasing the precipitation rate. Therefore, radar is not suitable for detection of seeding effects except <br />in special cases such as reported by Hobbs et al. (1981). Nevertheless, radar can be very valuable in those <br />cases with negligible natural snow, and can provide additional infonnation discussed below. <br /> <br />4.4 Some Final General Considerations <br /> <br />Even with a single seedline, natural variability can mask seeding effects. It is very important to not only <br />monitor the temporal changes at a target site, but spatial changes as well. This is best done by operating <br />some surface measurement stations in addition to the target. Such stations should be located crosswind <br />of the area to be affected by seeding so as to provide a record of natural variations with time. Radar <br />scanning can be very valuable in monitoring natural variations in cloud structure over the entire region <br />of the target. Such variations can mask, or be mistaken for, real seeding signatures and it is important <br />that they be documented. It is essential that enough observations be collected in both space and time to <br />detennine which perturbations are real seeding effects, which are simply natural variations, and which may <br />be seeding effects masked by natural variations. <br /> <br />Some of the past studies reviewed attempted to "piggy-back" physical experiments on what were basically <br />statistical designs. This was generally unsatisfactory. Most statistical experiments attempt to affect a <br />sizeable area for a significant time requiring seeding of a relatively large volume of aunosphere; for <br />example, the release of several seedlines for airborne seeding, Such seeding takes substantial time. Yet, <br />one of the main approaches for analyzing physical experiments is to examine temporal changes in expected <br />characterisncs (IPC, precipitation rate, degree of riming, etc.). This is best done by minimizing the time <br /> <br />38 <br />