WMOD00500 - Laserfiche WebLink

Home Browse Search

1134 JOURNAL OF APPLIED METEOROLOGY VOLUME 27 The total angles spanned by the instantaneous and wandering plumes at about the crest position are listed in Table 1 under "Inst" and "Span," respectively. Plume angles farther downwind sometimes differed. In general it was thought that if the variance in the wind direction was large, then the meander angle and instantaneous width of the plume would also be large, and that small wind direction variance would produce narrow plumes. The dependence of plume width on wind direction fluctuation has been documented in several studies (e.g., Slade 1968). When the winds were recorded at a seeding site, the wedge of the plume lo- cations was usually like a mirror image of the wedge of the wind vectors as plotted in Fig. 2. But when the plume angles were plotted against the standard devia- tion of wind direction from either the tower winds or the upper tethersonde winds, no significant correlation was found. There are several possible explanations. The plume angles were not measured by the same tech- niques that are used in air pollution studies; the con- centration spectrum across the plume was not mea- sured to determine standard deviation positions of the plume. Furthermore, Table 1 contains a mix of angles determined from both the superior ice particle plume and the inferior AgI plume measurements. As noted previously, the tower winds were 5 min averages from a vane with slow response, unlike the sensitive vanes used for measuring a true direction variation. The tethersonde balloon was itself the vane for its own di- rection measurements and responded rather slowly to changes. Furthermore, previous studies, such as those described in Slade (1968), were typically conducted over flat ground rather than over complex terrain like that of the mesa. The wind direction at a seeding site was usually found to be a good predictor of plume direction, and wind directions downwind offered guid- ance in determining later plume distortion, as is seen in Fig. 2. But a predictor for the plume spreading was not found in this study. The first part of Fig. 2 is devoted mostly to northerly flow cases. In Figs. 2a-2d the plume is seen to bend when it reaches the southwest arm of the mesa. In each case the winds in the notch (site C in Fig. 1) were gen- erally from the east, although sometimes with a large angular variance and light speeds. The angular variance of the winds on the northwest point (site D in Fig. 1) was so great in Figs. 2a and 2c that the wind vector wedges were nearly circular and were not plotted. The winds at site F in the upper right of Fig. 1 reversed during the experiment from NNW to S as evening drainage winds set in. Drainage winds are probably responsible for the final plume detection WSW of the seeding site, as indicated by the short dotted line away from the main plume wedge. The GJT winds near the times of the experiments in Figs. 2a and 2b were NNW near the level of the plume and were not a good indi- cator of the direction in which the plumes traveled. The experiments shown in Figs. 2c-2e are described in Super and Boe (1988). In these cases the plume edges were plotted from the ice particle data. Some- times the AgI plume was detected before the ice par- ticles because the generation of ice particles was de- pendent on the location of supercooled liquid water in the broken cloud deck, which contributed to the nar- rowness of the instantaneous plumes in Figs. 2c and 2e. As the winds became lighter in intensity (Fig. 2d) the plume broadened and distorted in shape. The experiment shown in Fig. 2f, the only one that was not conducted on the Grand Mesa, was designed to test the logistics and plume behavior associated with operating from near the crest of the Uncompahgre Pla- teau, since the Grand Mesa could presumably be seeded from such sites under southwest flow. These flows in- clude the wettest conditions over the Grand Mesa, and the greater upwind distance to the seeding site would allow more time for diffusion. The experiment occurred under westerly flow. The plume did not broaden with downwind distance, probably because there was very little directional variance in the upper winds ap- proaching the Uncompahgre Plateau. The near-surface winds measured by the tethersonde shifted from W to S during the experiment but with no noticeable effect on plume location or width. Either the southerly winds were immediately nullified by a counterflow at the crest of a local ridge or the plume rose rapidly to the level of westerly flow. The experiment in Fig. 2g occurred under highly sheared flow. Air approaching the mesa from the NW rose and created ice-saturated conditions near the crest of the Mesa. There it was met by SW flow coming over the top. Just above the aircraft sampling level there was an inversion, above which the winds reverted to NW flow. An ice fog plume was created by the seeding, and this plume traveled to the east. Ice nuclei were detected over a wide range of angles as the lower NW flow was stopped by the mesa crest and the air was caught in the SW flow coming over the top. Targeting of seeding materials in such a situation is obviously quite challenging. The existence of the ice particle plume in the otherwise clear air means that the AgI acted as deposition ice nuclei, or else froze a possible cloud of droplets that were produced in the plume by the water of combustion from the propane and acetone (Finnegan and Pitter 1987). There is no record of whether or not a visible cloud plume was emitted by the generator. The rest of the cases in Figs. 2h-2m show experi- ments in S to SW flow. In all cases the winds near the mesa were more southerly than in the Grand Junction soundings, in accordance with previous studies. In all but one case, either the IN or the ice particle plumes were detected downwind of the release site. The strong winds in Fig. 2i resulted in a lenticular cloud over the western arm of the mesa. The cloud position, based on photographs from the aircraft, is shown by the bold oval line. The AgI plume reached