During the past 2 decades, there has been increasing use of multielemental analytical methods to characterize airborne particulate matter and collected precipitation. With the availability of these large multivariate data sets and new multivariate statistical methods that have been made practical by the continuing increases in computer size and speed, there is an enhanced opportunity to extract information about atmospheric processes from data characterizing collected particles or precipitation. The original objective of this on-going research program was to continue the examination of the statistical problems associated with the source identification and mass apportionment of urban aerosols. During the current grant period, the program has changed its emphasis toward larger scale phenomena and the development and testing of statistical methods to extract information from analyzed samples obtained on a larger spatial and temporal scale. Specifically, the PI proposes to validate the Potential Source Contribution Function (PSCF) analysis and Total Potential Source Contribution Function (TPSCF) analysis Techniques both by applying PSCF to small-scale data sets and applying PSCF to the Across North America Tracer Experiment (ANATEX) data where the location and quantities of source emissions are known, to study and further develop three- way modeling via the application of existing methods to additional data sets and an investigation into alterations that may be necessary to these methods in the presence of constraints, and the further study of empirical orthogonal functions and their use in studying ozone transport.
