Laserfiche WebLink
<br />I <br /> <br />I <br /> <br />I <br /> <br />10-1 <br /> <br /> <br />I <br /> <br />10-4 <br /> <br />I <br /> <br />10-5 <br />o <br /> <br />2 3 4 5 <br />Drop Dlalllettr (111m) <br /> <br />6 <br /> <br />I <br />I <br /> <br />Fig. 5; Graphical representation of drop size <br />distributions for the natural and salt cases and <br />various rainwater contents (Ir). <br /> <br />The IAS modeling group under the leadership of <br />Dr. Orville has pioneered the use of bulk microphysical <br />models in hygroscopic seeding simulations. This effort <br />has evolved over time m~ch like that for ice-phase <br />seeding noted above. Initially, salt seeding was <br />simulated by reducing the intercept parameter of the <br />Marshall-Palmer distribution representing the raindrop <br />spectrum as in Myers and Orville (1972). Later <br />refinements by Orville and Kopp (1974) employed the <br />same initial reduction in the value of the intercept <br />parameter, but allowed it to gradually relax back to the <br />standard value by holding the slope of the distribution <br />constant after it had reached a certain level of maturity. <br />This is illustrated in Fig. 5. <br /> <br />I <br /> <br />I <br /> <br />I <br /> <br />I <br /> <br />The simple schemes described above can be <br />considered as first and second generation techniques for <br />simulating hygroscopic seeding. Third generation <br />techniques with an additional field variable representing <br />the seeding agent have been applied to hygroscopic <br />seeding during the current revival of interest. Details of <br />the techniques employed are given in Kopp (1994) and <br />Kopp et at. (1996). Two slightly different techniques <br />are used, depending on whether the seeding is with salt <br />particles dispensed by hoppers or the newer technology <br />of hygroscopic flares. In either case, the seeding agent <br />is introduced in the updraft region below cloud base <br />and subsequently transported by the flow field. <br /> <br />I <br /> <br />I <br /> <br />I <br /> <br />I <br /> <br />For conventional salt particle seeding, it is assumed <br />that rain (drizzle) is produced at a somewhat arbitrarily <br />determined rate for each gram of salt present at a grid <br />point. Cloud water is depleted in accordance with the <br />rain production. The salt is also assumed to have a <br />prescribed half-life so that it is not all converted at <br />once. Seeding rates of 100s of kg/km are simulated for <br />this formulation. <br /> <br />I <br /> <br />I <br /> <br />I <br /> <br />For hygroscopic flare seeding simulations, the <br />autoconversion process (conversion of cloud water to <br /> <br />I <br /> <br />rain) is activated wherever the seeding agent and cloud <br />water are collocated. The seeding agent is depleted as a <br />small percentage of the rate of rain water production by <br />autoconversion. Seeding rates are greatly reduced for <br />this technology with rates on the order of lOs to <br />I 008 g/km being typical. <br /> <br />Orville et al. (1998) offer an intriguingly simple <br />concept to test the possible effects of hygroscopic <br />seeding without simulating details of the process. They <br />reasoned that since the goal of hygroscopic seeding is <br />to transform a continental cloud naturally inefficient in <br />the warm rain process into a highly efficient maritime <br />cloud, the seeding potential could be determined by <br />comparing extreme situations. The basic assumption is <br />that a maritime case represents the maximum effect of <br />hygroscopic seeding on a continental convective cloud. <br />This tests two extreme situations, with the likely effect <br />of seeding being somewhere between the two extremes. <br /> <br />III. ICE-PHASE SEEDING SIMULATIONS <br /> <br />1. PRECIPITATION ENHANCEMENT <br /> <br />a. Warm Season Convective <br /> <br />Early efforts in the modeling of cloud seeding <br />effects by the IAS numerical modeling group were <br />directed at warm season convective clouds and was <br />done in conjunction with rain augmentation field <br />experiments conducted under Project Skywater of the <br />Bureau of Reclamation. This was in keeping with the <br />original mission of the IAS to promote weather <br />modification in the state of South Dakota. Much of the <br />initial emphasis involved applications of one- <br />dimensional steady-state models. Concurrent efforts <br />supported by NSF expanded the effort to one- and two- <br />dimensional time-dependent models which were also <br />undergoing a rapid evolution in microphysical <br />sophistication. The two-dimensional time-dependent <br />model became the standard tool in IAS cloud seeding <br />simulations in the early 1970s. <br /> <br />The starting point for realistic simulation of <br />summertime convective clouds in the Dakotas was the <br />microphysical parameterization scheme of Wisner et at. <br />(1972) with later improvements by Orville and Kopp <br />(1977). Early cloud seeding simulations using first <br />generation seeding techniques and relatively incomplete <br />microphysical schemes tended to produce coupled <br />seeding results, i.e. rain and hail would both experience <br />changes of the same sign as a result of seeding. With <br />the introduction of second generation seeding <br />techniques, we began to see results of cloud seeding <br />simulations which were no longer coupled in this way, <br />but still tended to suffer from poor representation of <br />important physical processes involved in cloud seeding. <br />Dramatic improvements were produced with the <br /> <br />19 <br />