You are here:
STATISTICAL CONSIDERATIONS IN THE EMPLOYMENT OF SAX (SCANNING ELECTRON MICROSOPY WITH AUTOMATED IMAGE ANALYSIS AND X-RAY ENERGY SPECTROSCOPY) RESULTS FOR RECEPTOR MODELS
Citation:
Johnson, D. AND J. Twist. STATISTICAL CONSIDERATIONS IN THE EMPLOYMENT OF SAX (SCANNING ELECTRON MICROSOPY WITH AUTOMATED IMAGE ANALYSIS AND X-RAY ENERGY SPECTROSCOPY) RESULTS FOR RECEPTOR MODELS. U.S. Environmental Protection Agency, Washington, D.C., EPA/600/D-83/075 (NTIS PB83225862), 1983.
Description:
Hundreds of thousands of individual particle measurements may be accumulated in a receptor model study employing Scanning electron microscopy with Automated image analysis and X-ray energy spectroscopy (SAX). At present, the summaries of these data are utilized in apportionment calculations. Two basic approaches have been taken to the sorting of these observations: variable algorithms which are sample specific, and fixed algorithms which have temporal and inter-airshed continuity. We report here the use of a 'fixed' sorting scheme developed for one area as it is applied to an airshed where no source specific information is available. Discriminant analysis, as applied to the individual particle characterizations, was shown to be very useful in improving the quality of the mathematical solution using a Q mode Factor/Vector analysis. It is concluded that statistical manipulation of the a 'fixed' sorting algorithm is shown to be feasible when unknown ambient aerosol samples are to be apportioned to sources on the basis of their comparison with library 'source signatures'.