Laserfiche WebLink
<br />at greater cost of computer resources. Additional studies are also possible <br />using finer resolution. Two studies were pr'oposed: (a) the sensitivity of dyna- <br />mics to the formulation of the microphysics; and (b) processes of entrainment/ <br />detrainment at the cloud boundary and internal mixing. <br /> <br />The sensitivity of the dynamical properties of the cloud to the microphysical <br />characteristics is directly related to the question of the degree to which the <br />dynamics of a cloud can be manipulated by changes in the microphysical pro- <br />cesses. Dynamical properties are believed to play the dominant role in cloud <br />and precipitation development. If the dynamics are influenced by microphysics, <br />this may be an effective way to effect changes in precipitation. Sensitivity <br />tests should include systematic changes in the rate of conversion from cloud <br />(non-precipitating liquid) to precipitating drops, the rate of conversion of <br />1 i qui d to ice and changes in the treatment of the fall of hydrometeors. <br />Accounting for the gravitational separation of hydrometeors has proven difficult <br />and the last topic should provide information on model sensitivity to this for- <br />mulation. It should also provide information on effects of changes brought <br />about by seeding because substantial differences in fall velocities of liquid <br />and ice hydrometeors commonly exist. <br /> <br />Debate continues concerning the location at which clear air from the environment <br />is entrained within clouds. The substantially uniform ("top hat") composition <br />of clouds in their early growing stage is not well matched by many cloud models, <br />apparently due to the coarse resolution employed and inadequate replication of <br />the internal mixing within clouds. The finer resolution experiments were recom- <br />mended to study the effects of different grid intervals on the simulation, the <br />importance of the resolution of different scales of motion, and in particular to <br />explore the possibility that the finer resolution will lead to new insight on <br />the entrainment and mixing processes. <br /> <br />2.2.2 Model architecture tests <br /> <br />The second type of sensitivity experiments concerned the mathematical architec- <br />ture of the model; these might include: (a) variations in numerical schemes <br />(for example: Smolarkiewicz1s scheme), (b) formulation of equations (for <br />example: entropy formulation, filtering methods, anelastic vs. non-anelastic, <br />etc.), and (c) focusing techniques, (for example: stretched or nested grids). <br /> <br />2.2.3 Experimental approach <br /> <br />The value of the sensitivity tests will be increased if the input data are com- <br />mon to another experiment being carried out for the workshop. Some of these <br />tests require substantial computer resources and it is likely that relatively <br />few participants would carry out these exper"iments. However, with input data <br />common to other aspects of the workshop, it Irli11 be possible to gain much more <br />benefit than if a unique data set were used. <br /> <br />Data sets for the sensitivity tests will be from days in which cumulus congest us <br />or small thunderstorms develop. The experimE~nt on the retention of the memory <br />of the perturbation will include two sets, one in which the outcome is strongly <br />dependent on the perturbation and another where it is not. The finer <br /> <br />5. <br />