THE RELATIONSHIP BETWEEN EMPIRICAL ORTHOGONAL FUNCTIONS AND SOURCES OF AIR-POLLUTION

Empirical Orthogonal Functions (EOFs) of airborne pollutant concentrations are shown to be closely related to the underlying advection differential equation. The average balance of sources and sinks of a pollutant is equal to the sum of the dot product of the gradient of the EOFs with specially weighted transport winds, plus some corrections for initial and final conditions. This relationship is the basis for the Source Identification Through Empirical Orthogonal Functions (SITEOF) model, a hybrid receptor model. This model is tested by applying it to simulated data from a simple dispersion model and 2 years of airborne fine sulfur concentrations from 24 western sites in the National Park Service monitoring network. In both cases, the predicted net source and sink areas derived form the SITEOF model are in good agreement with known major sulfur dioxide source areas. The effects of source location, turbulent diffusion, and averaging time of the observations on the model are discussed. Possible application include estimating emissions from fugitive or nontraditional sources, estimating production rates of secondary pollutants, and deposition rates.