You are here:
THE US EPA IMPLEMENTATION OF POSITIVE MATRIX FACTORIZATION AND A NEW APPROACH TO UNCERTAINTY EMISSIONS
Citation:
HOPKE, P. K., P. PAATERO, AND S. I. EBERLY. THE US EPA IMPLEMENTATION OF POSITIVE MATRIX FACTORIZATION AND A NEW APPROACH TO UNCERTAINTY EMISSIONS. Presented at International Environmental Modeling and Software Society , Burlington, VT, July 09 - 13, 2006.
Impact/Purpose:
The goal of this task is to develop methods and models to reduce the uncertainty in quantifying local and regional air pollutant source impacts on ambient samples collected in speciated PM, air toxic, and semi-continuous measurement networks. A combination of high resolution sampling, organic and inorganic analytical methods, and models will be developed and evaluated to reduce the uncertainty in source apportionment:
(1) semi-continuous inorganic species sampling
(2) inorganic analysis
(3) organic analysis for medium flow samples
(4) multivariate receptor models for ambient samples
(5) regional and local models
In addition, this task contributes to two additional tasks that have research focused on reducing the uncertainty in source apportionment: Identify Sources of Human Exposure (21176), and NAAQS implementation (21179).
Description:
This abstract describes the approach implemented in EPA's version of Positive Matrix Factorization (EPA PMF) to estimate uncertainties in the modeled solutions. Details are provided regarding sources of uncertainty in constrained factor analytic models and how these approaches are addressed with bootstrapping and pulling, the techniques used in EPA PMF. Results using simulated and real data are presented.