WMOD00508 - Laserfiche WebLink

Home Browse Search

292 JOURNAL OF APPLIED METEOROLOGY VOLUME 29 chopper on the aircraft which produced pellets 5 to 20 mm in diameter. Clear air tests indicated that the CO2 formed a vertical curtain extending 700 m below the aircraft. Although initially one successful experiment was conducted with 20 g AgI pyrotechnic flares (Stewart and Marwitz 1982), difficulties that appeared in later years, coupled with the expense of the flares and the warm temperatures at which seeding was required, nearly precluded their use. Clear air tests indicated that in some batches less than half of the flares burned for the 30 s required. Activity tests in a cloud chamber indicated reduced yield of ice nuclei by two to three orders of magnitude compared to AgI burned in ace- tone, particularly for flares with the tracer cesium added (DeMott and Grant 1986). Thus to ensure glaciation in clouds seeded at temperatures> -lOoC, CO2 was believed to be the best material. During the last year of the experiment the seeding material was changed to an AgI NH41 NH4Cl04 mix- ture burned in an acetone solution. This was done for several reasons. The seeding material is economical and easier to handle than CO2. The delivery system is simple, and the dispensing of seeding material can be directly confirmed by measuring the flame temperature of the acetone burner. Seeding can be conducted for long periods of time with a small amount of material, and finally, the AgI can be used as a tracer since it can be measured with an ice nucleus counter. The AgI was released at a rate of 0.4 g km -1 using a generator con- taining a 3 percent by weight solution of AgI NH41 NH4Cl04 with the ammonium perchlorate at a 30 mole ratio. This solution was chosen since tests indicated it would produce approximately 10 13 nuclei g-l at -60C (DeMott et al. 1983). With this activity a rough cal- culation of the ICC resulting from seeding can be made. Assuming that the AgI "cylinder" expands at 1.0 m S-1 and that 80 percent of the ice nuclei have activated (DeMott et aI. 1983), the ICC would be approximately 10 L -1 after 10 min. However, if contact nucleation is the primary nucleation mechanism for AgI NH41, as DeMott et al. suggest, then this estimate is too high. Including the rate at which ice nuclei are scavenged by cloud droplets (Slinn 1971), and assuming 100 cm -3 of5 11m cloud drops, instead of the 2000 to 4000 cm-3 in the cloud chamber where the material was tested, the expected ICC would be <1.0 L -1 after 10 min. If this is the case, increases in ICC due to seeding would not be measurable. 3. Case studies The following two case studies present results from one placebo experiment and two experiments that demonstrate seeding effects. The first case study in- cludes the placebo experiment and a seeding experi- ment using AgI NH41 NH4Cl04. The second case study is a seeding experiment using dry ice on a day when the cloud was too shallow for the research aircraft to conduct its normal measurement profile. To infer seeding effects from aircraft data requires knowing coincidence times between aircraft and seed- line position. To accomplish this, seedlines were ad- vected downwind using aircraft measured winds, or the aircraft pointer data, and were then compared with the aircraft flight path to determine intersection times. Data from the particle probes on the aircraft were then examined for ice crystal plumes at the intersection times. Peaks in the ICC data were expected for seeded cases. The largest source of error in this technique oc- curs from the use of a constant advection speed for the seedlines; however, an error of 2 m s -I results in a displacement of only 3.6 km in seedline position after 30 min, which corresponds to < 1 min of aircraft data. Thus, this technique should provide a good indication of where to expect seeding effects in the aircraft data. When AgI was used seedline penetration could be confirmed directly using the NCAR ice nucleus counter on the aircraft. Empirical tests of the NCAR counter indicated a 25-40 s delay between intake of ice nuclei and registration of the particles. This delay is consistent with Super and Boe ( 1988). The ice nucleus concen- trations to be presented here only account for sampling volume. Loss to sidewalls and other problems noted by Langer (1973) and Sackiw et al. (1984) were not considered. '1 I a. 18 December 1986 1) CLOUD STRUCTURE AND ORGANIZATION On the morning of 18 December a weak short-wave trough moved into a split in the upper-level flow over California. The trough weakened considerably by 0000 (all times UTC) 19 December 1986. Preceding the as- sociated surface front was an upper-level cold surge, of the type described by Hobbs (1978), that passed SHR just after the 1500 sounding, Fig. 2. Satellite im- ages (not shown) indicated that a cloud band moved through the Sacramento Valley with the cold surge. The precipitation associated with this cold surge moved over the American River basin between 1430 and 1630. The surface frontal band passed SHR between 1800 and 2100, evidenced by the substantial warming aloft from subsidence behind the front. The frontal band could not be characterized well in the lowest 1400 m, although an increase in precipitation marked passage over KGV between 2300 and 0100. Lack of a strong thermal discontinuity (Fig. 2) and the 4 to 5 hours required to move from SHR to KGV indicate the weakness of the front. Two fixed-target experiments were conducted between 1800 and 2000, after passage of the upper-level cold surge, but prior to passage of the surface cold front. This has been noted by Reynolds and Kuciauskas ( 1988) to be a favonlble location for the development ofSLW. The ascent sounding and MOCA descent by the