Assessing the Lagrangian Framework and State Estimation for Forward and Inverse Atmospheric Transport and Dispersion Modeling

Open Access
Annunzio, Andrew Joseph
Graduate Program:
Doctor of Philosophy
Document Type:
Date of Defense:
March 01, 2011
Committee Members:
  • Sue Ellen Haupt, Committee Chair
  • George Young, Committee Chair
  • Hampton Shirer, Committee Member
  • Diane Henderson, Committee Member
  • Leonard Joel Peltier, Committee Member
  • Paul Bieringer, Committee Member
  • Entity
  • State Estimation
  • Transport and Dispersion
  • Source Term Estimation
  • FFT 07
In atmospheric transport and dispersion modeling, there are two frameworks used to predict the evolution of an airborne contaminant: Eulerian and Lagrangian. Each framework has its advantages and disadvantages. An advantage of the Eulerian or field framework is that the governing variables are space filling, and describe the entire concentration and velocity field. Thus, the evolution of these variables are governed by a coupled system of partial differential equations and an equation of state. Although these equations in their full form can be solved only by numerical approximation, describing flow field evolution by a set of equations allows one to grasp how forces change flow field properties at specific points. The Lagrangian framework, in contrast, concentrates on entity state evolution where the superposition of such entities describes the concentration field. An advantage of the Lagrangian framework is that under certain circumstances, both the evolution of the flow and the concentration field can be described by a system of ordinary or partial differential equations that determine entity state evolution. These equations can be simpler than the Eulerian counterpart. This is illustrated in the thesis for both forward and inverse atmospheric transport and dispersion modeling. Specifically, a Lagrangian source term estimation algorithm is developed to determine the source location of a contaminant. It is shown that this method performs well when tested on the FFT 07 dataset for single and multiple contaminant releases.