WMOD00289 ‎(2)‎

• Title: Observations and Numerical Simulations of Precipitation Mechanisms in Natural and Seeded Convective Clouds
• Prepared By: Lroen D. Nelson
• Date: 3/1/1979
• State: IL
• Weather Modification - Doc Type: Report

30

A potentially more serious model pseudodiffusion error may occur due to

the use of a variant of upstream differencing (see Kovetz and Olund, 1969), for

particulate mass advection (condensation/sublimation). In this model, condensation/

sublimation is done in series with the other parallel model processes to permit the

continued recalculation of a maximally large U d~t/~ via a separate cClnden-
con

sational At, thus minimizing pseudodiffusive errors in particulate mass growl'h. Since

the J and K scale particulate spacings ore far apart in mass at larger sizes, it might

be thought that the numerical diffusion errors in condensation would become! quite

large at larger sizes. Fortunately, this effect is offset by the rapid decreasE~ with

particle size of the effective condensation driving force.

Because of the complicating effects of the nonlinear mass scale and '~he non-

constant advective velocity (mass condensation rate) involved in numerical solution

of the condensation equation by the upstream difference technique, the moss

pseudodiffusion effect of the numerical scheme is not easily solved for analytically.

Silverman and Gloss (1973) have investigated this problem in detail by a series of

model runs using different numbers (mass spacings) of hydrometeor mass classE~s. They

conclude that 45 log scale mass classes to cover the hydrometeor size spectrum (as

is done in this model for both ice and liquid particles) leads to reasonable and

stable results while models using appreciably fewer size classes will be subject to

significant particle size pseudodiffusion errors.

As a check on the applicability of their conclusions to this larger model in-

c1uding ice phase effects, the model simulation of the August 14, 1975 Big Spring

Texas Cloud (given later as Table 4 and Figures 20 through 26) was repeated with