WMOD00473 - Laserfiche WebLink

Home Browse Search

1992 JOURNAL OF CLIMATE AND APPLIED METEOROLOGY VOLUME 22 upward motion approaching the barrier. Crystal con- centrations of 5-20 L -1 were found upwind of the barrier with almost no temperature dependence. Rauber and Grant (1981 b), also present model results in good agreement with their observations. More recent observations from several more storms gave further evidence that most of the liquid water was located over the crest and upper slopes of the Park Range (Rauber et aI., 1982). Hill (1980a), in discussing winter orographic storms in northern Utah, noted that, "our data indicate that the production of supercooled water is closely asso- ciated with vertical air motion which in turn is found primarily at or near mountain-top levels". Thus, it appears that liquid water concentrations may generally be quite low in stable orographic clouds. The only region of significant liquid water may be within ~.1 kIn of the ground above the upwind slope of the barrier, over the ridge and perhaps for a short distance beyond. This likely corresponds to the region of most pro- nounced upward motion. In this situation, ice crystal growth may be very limited outside the zone of su- percooled liquid water. Simple calculations of diffu- sional growth and. subsequent ice crystal trajectories suggest that crystals formed at elevations well above the liquid water zone (e.g. near cloud top) are generally carried beyond a barrier the width of the Bridger Range without settling to the surface. The crystals most im- portant to the snowfall processes may be those formed well below cloud top that are carried quasi-horizontally into the liquid water zone where rapid growth can occur. While understanding of both natural and seeded snowfall process is still incomplete, the following gen- eral statements about artificial seeding can be made. In order for cloud seeding to increase snowfall from winter clouds over mountainous terrain, several links in a physical chain of events must exist. First, seeding material must be successfully and reliably produced. Second, this material must be transported into a region of cloud that has supercooled water or ice supersat- uration in excess of that which can be converted to ice by naturally produced ice crystals. Third, the seeding material must have dispersed sufficiently upon reaching this region so that a significant volume is affected by the desired concentration range of ice nuclei or the resulting ice crystals. In the case of AgI seeding this requires, fourth, that the temperature be low enough for substantial nucleation to occur. Once ice crystals form, they must remain in an environment suitable for growth long enough to enable fallout to occur, generally prior to their being carried beyond the mountain barrier where downslope motion, cloud evaporation and ice crystal sublimation typically exist. The BRE was not able to acquire direct physical evidence that each of these steps in the physical process occurred in a significant fraction of the orographic storms. Only AgI generator operation, precipitation measured by the gage network, certain other surface observations and upper air parameters were routinely monitored. However, the BRE did obtain evidence that the first four steps frequently occurred. Consid- erable uncertainties still remain concerning the entire physical process, particularly with regard to the growth and trajectories of ice particles. This post hoc statistical analysis is an attempt to reduce these uncertainties. 3. Experimental equipment a. Precipitation gage network A network of 31 precipitation monitoring sites was maintained during the 1970-71 and 1971-72 winters and is shown in Fig. 1. Twenty-eight of these sites were east of the Main Ridge, one was 28 km south of the ,southern seeding site and the other two were at the seeding sites. These latter three were used as control gages in the statistical analysis herein. The mechanisms of standard Belfort weighing gages were used throughout the network. However, special gage shells were constructed with a 28.7 cm diameter orifice. This provided twice the sampling area of the standard 8 inch gage, which prevented "capping" by snow buildup, and dm.ibled resolution on the gage chart. Each gage was equipped with an Alter-II type windshield (Warnick, 1956). All gages were positioned in relation to local terrain features or forest cover so as to minimize wind effects. The higher elevation sites, including the control gages and gages on the Bangtail Ridge, were located in natural clearings in the forest. These would be classified as "overprotected" by Brown and Peck (1962). It is not known how well measurements in such sites represent the absolute snowfall amount over the larger, mostly forested area. However, as shown in Part I, comparisons between gages and snowboards showed excellent agreement within such sites (r = 0.99) and correlation coefficients of approximately 0.95 were common for 24 h totals between sheltered gage sites located ~4 km apart. Thus, at.least in a relative sense, the gages appeared to provide reliable estimates of precipitation. All precipitation charts were manually reduced to the nearest 0.01 inch by two independent teams of data clerks. Any differences were then reconciled by a third examination. Approximately 96% of all possible precipitation data was retrieved from the gage charts. About three-fourths of the remaining 4% had total pr~cipitation known but time distribution unknown, usually due to clock stoppages. All missing data.were estimated with the aid of a series of maps, similar to Fig. 1, which were prepared for all consecutive 6 h periods with missing data. All available precipitation totals were noted by each gage site, and information on "amounts known but time distribution unknown" was also noted. However, each map was coded and nothing on it indicated the date in question, whether the day was seeded, or even if the period was from an