AAAR 32nd Annual Conference
September 30 - October 4, 2013
Oregon Convention Center
Portland, Oregon, USA
Abstract View
Aerosol Dynamics Simulation Using Sparse Particle Methods
ROBERT MCGRAW, Brookhaven National Laboratory
Abstract Number: 422 Working Group: Aerosols, Clouds, and Climate
Abstract A new class of sparse particle representations for an aerosol population is presented. As defined here, "sparse particle representation" refers to the replacement of a typically continuous particle distribution function (pdf) by a small set of weighted delta functions in particle size and/or composition coordinates. Specifically, class members are defined here as solutions to a constrained optimization problem involving maximization or minimization of an aerosol property using linear programming (LP) methods. The quadrature method of moments (QMOM), in which a set of lower-order moments is evolved by mapping them to an equivalent set of quadrature abscissas and weights for closure, is shown to fall into this class. The new LP-based approach allows for a much broader class of aerosol properties to be used to generate the sparse particle distributions; whereas the QMOM is limited to moments. The new sparse distributions are shown to provide a set of abscissas and weights that can be used both for quadrature approximation of aerosol physical and optical properties and for closure of the dynamical equations governing evolution of the pdf. In the first case, approximation of aerosol properties, nested bounds on the property are obtained through the sequential addition of either model- or observation-based constraints, with the most relevant measurements providing the greatest reduction in uncertainty, quantified in LP as a reduction in the size of the feasible solution set. In the second case, aerosol dynamics, a generalization of the Jacobian matrix transformation (JMT) [McGraw and Wright, Aerosol Science 34, 189-209 (2003)] enables a closed set of equations for aerosol property evolution to be obtained directly in terms of the sparse pdf. As a result, JMT closure is extended to a much broader set of aerosol properties than moments.