References

ACPD

Atmospheric Chemistry and Physics Discussions

ACPD

1680-7375

Copernicus GmbH

Göttingen, Germany

10.5194/acpd-8-2977-2008

Assessing positive matrix factorization model fit: a new method to estimate uncertainty and bias in factor contributions at the daily time scale

Hemann

J. G.

¹ Brinkman

G. L.

² Dutton

S. J.

² Hannigan

M. P.

² Milford

J. B.

² Miller

S. L.

Department of Applied Mathematics, University of Colorado, Boulder, USA

Department of Mechanical Engineering, University of Colorado, Boulder, USA

14 02 2008

8 1 2977 3026

This is an open-access article ditributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

This article is available from http://www.atmos-chem-phys-discuss.net/8/2977/2008/acpd-8-2977-2008.html

The full text article is available as a PDF file from http://www.atmos-chem-phys-discuss.net/8/2977/2008/acpd-8-2977-2008.pdf

A Positive Matrix Factorization receptor model for aerosol pollution source apportionment was fit to a synthetic dataset simulating one year of daily measurements of ambient PM<sub>2.5</sub> concentrations, comprised of 39 chemical species from nine pollutant sources. A novel method was developed to estimate model fit uncertainty and bias at the daily time scale, as related to factor contributions. A balanced bootstrap is used to create replicate datasets, with the same model then fit to the data. Neural networks are trained to classify factors based upon chemical profiles, as opposed to correlating contribution time series, and this classification is used to align factor orderings across results associated with the replicate datasets. Factor contribution uncertainty is assessed from the distribution of results associated with each factor. Comparing modeled factors with input factors used to create the synthetic data assesses bias. The results indicate that variability in factor contribution estimates does not necessarily encompass model error: contribution estimates can have small associated variability yet also be very biased. These results are likely dependent on characteristics of the data.

References 1

Anderson, M. J., Daly, E. P., Miller, S. L., and Milford, J. B.: Source apportionment of exposures to volatile organic compounds: II. Application of receptor models to TEAM study data, Atmos. Environ., 36, 3643–3658, 2002.

Brinkman, G., Vance, G., Hannigan, M. P., and Milford, J. B.: Use of synthetic data to evaluate positive matrix factorization as a source apportionment tool for PM2.5 exposure data, Environ. Sci. Technol., 40, 1892–1901, 2006.

Cadle, S. H., Mulawa, P., Hunsanger, E. C., Nelson, K., Ragazzi, R. A., Barrett, R., Gallagher, G. L., Lawson, D. R., Knapp, K. T., and Snow, R.: Light-duty motor vehicle exhaust particulate matter measurement in the Denver, Colorado, area, Journal of the Air and Waste Management Association, 49, 164–174, 1999.

Carlstein, E.: The use of subseries values for estimating the variance of a general statistic from a stationary sequence, Ann. Stat., 14, 1171–1179, 1986.

Chatfield, C.: Model uncertainty, data mining and statistical inference, Journal of the Royal Statistical Society Series A-Statistics in Society, 158, 419–466, 1995.

Chen, L. W. A., Watson, J. G., Chow, J. C., and Magliano, K. L.: Quantifying PM2.5 source contributions for the San Joaquin valley with multivariate receptor models, Environ. Sci. Technol., 41, 2818–2826, 2007.

Chinkin, L. R., Coe, D. L., Funk, T. H., Hafner, H. R., Roberts, P. T., Ryan, P. A., and Lawson, D. R.: Weekday versus weekend activity patterns for ozone precursor emissions in California's south coast air basin, Journal of the Air and Waste Management Association, 53, 829–843, 2003.

Davison, A. C., Hinkley, D. V., and Schechtman, E.: Efficient bootstrap simulation, Biometrika, 73, 555–566, 1986.

Dockery, D. W., Pope, C. A., Xu, X. P., Spengler, J. D., Ware, J. H., Fay, M. E., Ferris, B. G., and Speizer, F. E.: An association between air-pollution and mortality in 6 United-States cities, New England Journal of Medicine, 329, 1753–1759, 1993.

Dominici, F., McDermott, A., Daniels, M., Zeger, S. L., and Samet, J. M.: Revised analyses of the national morbidity, mortality, and air pollution study: Mortality among residents of 90 cities, Journal of Toxicology and Environmental Health, Part A, 68, 1071–1092, 2005.

Eberly, S. I.: EPA PMF 1.1 User's Guide, U.S. Environmental Protection Agency, Research Triangle Park, NC, 2005.

Efron, B.: 1977 Rietz lecture - bootstrap methods - another look at the jackknife, Ann. Stat., 7, 1–26, 1979.

Fine, P. M., Cass, G. R., and Simoneit, B. R. T.: Chemical characterization of fine particle emissions from the fireplace combustion of wood types grown in the midwestern and western United States, Environ. Eng. Sci., 21, 387–409, 2004.

Gleason, J. R.: Algorithms for balanced bootstrap simulations, Am. Stat., 42, 263–266, 1988.

Hall, P., Horowitz, J. L., and Jing, B. Y.: On blocking rules for the bootstrap with dependent data, Biometrika, 82, 561–574, 1995.

Hannigan, M. P.: Mutagenic particulate matter in air pollutant source emissions and in ambient air, Doctorate of Philosophy, California Institute of Technology, Pasadena, CA, 221 pp., 1997.

Haykin, S.: Neural networks: A comprehensive foundation, Prentice Hall; 2nd edition, Upper Saddle River, 1998.

Heidam, N. Z.: Bootstrap estimates of factor model variability, Atmos. Environ., 21, 1203–1217, 1987.

Huang, S. L., Rahn, K. A., and Arimoto, R.: Testing and optimizing two factor-analysis techniques on aerosol at Narragansett, Rhode Island, Atmos. Environ., 33, 2169–2185, 1999.

Kiefer, N. M. and Vogelsang, T. J.: A new asymptotic theory for heteroskedasticity-autocorrelation robust tests, Economet. Theor., 21, 1130–1164, 2005.

Kim, E., and Hopke, P. K.: Source identifications of airborne fine particles using positive matrix factorization and US Environmental Protection Agency positive matrix factorization, Journal of the Air and Waste Management Association, 57, 811–819, 2007.

Kunsch, H. R.: The jackknife and the bootstrap for general stationary observations, Ann. Stat., 17, 1217–1241, 1989.

Lahiri, S. N.: Effects of block lengths on the validity of block resampling methods, Probability Theory and Related Fields, 121, 73–97, 2001.

Larsen, R. K., and Baker, J. E.: Source apportionment of polycyclic aromatic hydrocarbons in the urban atmosphere: A comparison of three methods, Environm. Sci. Technol., 37, 1873–1881, 2003.

Lee, E., Chan, C. K., and Paatero, P.: Application of positive matrix factorization in source apportionment of particulate pollutants in Hong Kong, Atmos. Environ., 33, 3201–3212, 1999.

Lewis, C. W., Norris, G. A., Conner, T. L., and Henry, R. C.: Source apportionment of Phoenix PM2.5 aerosol with the Unmix receptor model, Journal of the Air and Waste Management Association, 53, 325–338, 2003.

Lough, G. C.: Sources of metals and NMHCs from motor vehicle roadways, Doctor of Philosophy University of Wisconsin, Madison, Madison, WI, 2004.

Munakata, T.: Fundamentals of the new artifical intelligence: Beyond traditional paradigms, Graduate texts in computer science, edited by: Grles, D., and Schneider, F. B., Springer-Verlag, New York, 1998.

Nitta, H., Ichikawa, M., Sato, M., Konishi, S., and Ono, M.: A new approach based on a covariance structure model to source apportionment of indoor fine particles in Tokyo, Atmos. Environ., 28, 631–636, 1994.

Norris, G.A., Vedantham, R., and Duvall, R.: EPA Unmix 6.0 Fundamentals and User Guide; U.S. Environmental Protection Agency, Research Triangle Park, NC, 2007.

Olson, D. A., Norris, G. A., Seila, R. L., Landis, M. S., and Vette, A. F.: Chemical characterization of volatile organic compounds near the World Trade Center: Ambient concentrations and source apportionment, Atmos. Environ., 41, 5673–5683, 2007.

Paatero, P. and Tapper, U.: Positive matrix factorization - a nonnegative factor model with optimal utilization of error-estimates of data values, Environmetrics, 5, 111–126, 1994.

Paatero, P., Hopke, P. K., Song, X. H., and Ramadan, Z.: Understanding and controlling rotations in factor analytic models, Chemometrics and Intelligent Laboratory Systems, 60, 253–264, 2002.

Paatero, P., Hopke, P. K., Begum, B. A., and Biswas, S. K.: A graphical diagnostic method for assessing the rotation in factor analytical models of atmospheric pollution, Atmos. Environ., 39, 193–201, 2005.

Peel, J. L., Tolbert, P. E., Klein, M., Metzger, K. B., Flanders, W. D., Todd, K., Mulholland, J. A., Ryan, P. B., and Frumkin, H.: Ambient air pollution and respiratory emergency department visits, Epidemiology, 16, 164–174, 2005.

Polissar, A. V., Hopke, P. K., and Paatero, P.: Atmospheric aerosol over Alaska - 2. Elemental composition and sources, J. Geophys. Res.-Atmos., 103, 19 045–19 057, 1998.

Politis, D. N. and White, H.: Automatic block-length selection for the dependent bootstrap, Economet. Rev., 23, 53–70, 2004.

Pope, C. A., Burnett, R. T., Thun, M. J., Calle, E. E., Krewski, D., Ito, K., and Thurston, G. D.: Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution, J. Am. Med. Assoc., 287, 1132–1141, 2002.

Ramadan, Z., Song, X. H., and Hopke, P. K.: Identification of sources of Phoenix aerosol by positive matrix factorization, Journal of the Air and Waste Management Association, 50, 1308–1320, 2000.

Rogge, W. F., Hildemann, L. M., Mazurek, M. A., Cass, G. R., and Simoneit, B. R. T.: Sources of fine organic aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks, Environmental Science and Technology, 27, 636–651, 1993.

Rogge, W. F., Hildemann, L. M., Mazurek, M. A., Cass, G. R., and Simoneit, B. R. T.: Sources of fine organic aerosol. 3. Road dust, tire debris, and organometallic brake lining dust - roads as sources and sinks, Environmental Science and Technology, 27, 1892–1904, 1993.

Rogge, W. F., Hildemann, L. M., Mazurek, M. A., Cass, G. R., and Simoneit, B. R. T.: Sources of fine organic aerosol. 4. Particulate abrasion products from leaf surfaces of urban plants, Environ. Sci. Technol., 27, 2700–2711, 1993.

Rogge, W. F., Hildemann, L. M., Mazurek, M. A., Cass, G. R., and Simoneit, B. R. T.: Sources of fine organic aerosol. 5. Natural-gas home appliances, Environ. Sci. Technol., 27, 2736–2744, 1993.

Rumelhart, D. E., Hinton, G. E., and Williams, R. J.: Learning representations by back-propagating errors, Nature, 323, 533–536, 1986.

Schauer, J. J.: Source contributions to atmospheric organic compound concentrations: Emissions measurements and model predictions, Doctor of Philosophy, California Institute of Technology, Pasedena, CA, 399 pp., 1998.

Schauer, J. J., Kleeman, M. J., Cass, G. R., and Simoneit, B. R. T.: Measurement of emissions from air pollution sources. 1. C-1 through C-29 organic compounds from meat charbroiling, Environ. Sci. Technol., 33, 1566–1577, 1999.

Shafri, H. Z. M., Suhaili, A., and Mansor, S.: The performance of maximum likelihood, spectral angle mapper, neural network and decision tree classifiers in hyperspectral image analysis, Journal of Computer Science, 3, 419–423, 2007.

Singh, K.: On the asymptotic accuracy of Efron's bootstrap, Ann. Stat., 9, 1187–1195, 1981.

Tong, C. S. and Cheng, K. C.: Mass spectral search method using the neural network approach, Chemometrics and Intelligent Laboratory Systems, 49, 135–150, 1999.

van der Meer, F.: The effectiveness of spectral similarity measures for the analysis of hyperspectral imagery, Int. J. Appl. Earth Obs., 8, 3–17, 2006.

Watson, J. G., Fujita, E., Chow, J., Zielinska, B., Richards, L. W., Neff, W., and Dietrich, D.: Northern Front Range Air Quality Study final report, Desert Research Institute, Fort Collins, CO, 1998.

Werbos, P. J.: Beyond regression: New tools for prediction and analysis in the behavioral sciences, Doctorate of Philosophy, Harvard University, Cambridge, 1974.