HR: 1340h
AN: A23D-1571    [Abstracts]
TI: Errors of Nonlinear Advection Algorithms and Their Implications for Studying the Aerosol- Cloud Interaction
AU: * Ovtchinnikov, M
EM: mikhail@pnl.gov
AF: Pacific Northwest National Laboratory, P.O. Box 999, Richland, WA 99352, United States
AU: Easter, R C
EM: Richard.Easter@pnl.gov
AF: Pacific Northwest National Laboratory, P.O. Box 999, Richland, WA 99352, United States
AU: Wang, W
EM: Weiguo.Wang@pnl.gov
AF: Pacific Northwest National Laboratory, P.O. Box 999, Richland, WA 99352, United States
AB: Monotonic numerical schemes are commonly used in modeling tracer transport in the atmosphere to prevent the appearance of non-physical negative values and various flux correction procedures are applied to preserve sharp gradients in the simulated fields. Because these monotonicity and flux correction constrains are nonlinear, the monotonicity for individual species does not guarantee monotonicity for their sum or other linear combination. As a result, artificial (numerical) redistribution of total aerosol number concentration in a 3D model, for example, can mask the effects of real aerosol processing by clouds. The problem is potentially relevant to any model that relies on preserving certain relations between independently advected variables (e.g., multi-moment aerosol/cloud modules, sectional microphysics, multicomponent chemistry). This study quantifies the errors in three widely used advection algorithms by extending traditional single-shape advection tests to a multiple shapes and multiple species with prescribed linear sum. A normalization procedure that preserves positive definiteness of individual variables and monotonicity of their linear combination, such as the total number concentration of aerosol particles, is also discussed.
DE: 0368 Troposphere: constituent transport and chemistry
DE: 0545 Modeling (4255)
DE: 0560 Numerical solutions (4255)
SC: Atmospheric Sciences [A]
MN: 2007 Fall Meeting