Initial Exploration of Advanced Data Analysis Methods to Assist Air Quality Management;Final rept

摘要

This project demonstrated advanced statistical methods for analyzing aerosol data. The first test applied Positive Matrix Factorization (PMF) to IMPROVE aerosol data from Crater Lake National Park (CRLA), Lassen Volcanic National Park (LAVO), and San Gorgonio National Wilderness (SAGO). For CRLA and LAVO the objective was to separate and quantify the influence of Asian dust. At these sites, six common 'sources' were identified, all exhibiting similar chemical compositions and seasonal variations. The source profile of Asian Dust agreed well with other Asian Dust studies, and was strongly correlated with secondary sulfate. The SAGO IMPROVE site is downwind of Los Angeles. There the objective was to distinguish aerosols from gasoline vs. diesel emissions. The analysis produced diesel and gasoline profiles based on OC/EC ratios. SAGO also showed the Asian dust factor seen at CRLA and LAVO. The second task applied advanced methods to Aerosol Time-of-Flight Mass Spectrometry (ATOFMS) data from Fresno to distinguish diesel/gasoline emissions, develop a calibration model based on ART-2a and PLS, and estimate uncertainties for ATOFMS data. PMF results, however, were unclear, suggesting the need for future collaboration with the ATOFMS team of Prof. Prather of the University of California, San Diego.