A novel probabilistic source apportionment approach: Bayesian auto-correlated matrix factorization
<p>The concentrations of atmospheric particulate matter and many of its constituents are temporally auto-correlated. However, this information has not been utilized in source apportionment methods. Here, we present a Bayesian matrix factorization model (BAMF) that considers the temporal auto-c...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2024-02-01
|
| Series: | Atmospheric Measurement Techniques |
| Online Access: | https://amt.copernicus.org/articles/17/1251/2024/amt-17-1251-2024.pdf |
| _version_ | 1827319111920648192 |
|---|---|
| author | A. Rusanen A. Rusanen A. Björklund M. I. Manousakas J. Jiang J. Jiang M. T. Kulmala M. T. Kulmala M. T. Kulmala K. Puolamäki K. Puolamäki K. R. Daellenbach |
| author_facet | A. Rusanen A. Rusanen A. Björklund M. I. Manousakas J. Jiang J. Jiang M. T. Kulmala M. T. Kulmala M. T. Kulmala K. Puolamäki K. Puolamäki K. R. Daellenbach |
| author_sort | A. Rusanen |
| collection | DOAJ |
| description | <p>The concentrations of atmospheric particulate matter and many of its constituents are temporally auto-correlated. However, this information has not been utilized in source apportionment methods. Here, we present a Bayesian matrix factorization model (BAMF) that considers the temporal auto-correlation of the components (sources) and provides a direct error estimation. The performance of BAMF is compared with positive matrix factorization (PMF) using synthetic Time-of-Flight Aerosol Chemical Speciation Monitor data, representing different urban environments from typical European towns to megacities. We find that BAMF resolves sources with overall higher factorization performance (temporal behavior and bias) than PMF on all datasets with temporally auto-correlated components. Highly correlated components continue to be challenging and ancillary information is still required to reach good factorizations. However, we demonstrate that adding even partial prior information about the chemical composition of the components to BAMF improves the factorization. Overall, BAMF-type models are promising tools for source apportionment and merit further research.</p> |
| first_indexed | 2024-03-07T13:58:11Z |
| format | Article |
| id | doaj.art-822e99b53654480ebc75d84dbc523830 |
| institution | Directory Open Access Journal |
| issn | 1867-1381 1867-8548 |
| language | English |
| last_indexed | 2024-04-25T00:14:38Z |
| publishDate | 2024-02-01 |
| publisher | Copernicus Publications |
| record_format | Article |
| series | Atmospheric Measurement Techniques |
| spelling | doaj.art-822e99b53654480ebc75d84dbc5238302024-03-13T05:46:25ZengCopernicus PublicationsAtmospheric Measurement Techniques1867-13811867-85482024-02-01171251127710.5194/amt-17-1251-2024A novel probabilistic source apportionment approach: Bayesian auto-correlated matrix factorizationA. Rusanen0A. Rusanen1A. Björklund2M. I. Manousakas3J. Jiang4J. Jiang5M. T. Kulmala6M. T. Kulmala7M. T. Kulmala8K. Puolamäki9K. Puolamäki10K. R. Daellenbach11Institute for Atmospheric and Earth System Research (INAR)/Physics, Faculty of Science, University of Helsinki, 00014 University of Helsinki, FinlandAtmospheric Composition Research, Finnish Meteorological Institute, 00101 Helsinki, FinlandDepartment of Computer Science, Faculty of Science, University of Helsinki, 00014 University of Helsinki, FinlandLaboratory of Atmospheric Chemistry, Paul Scherrer Institute (PSI), 5232 Villigen-PSI, SwitzerlandLaboratory of Atmospheric Chemistry, Paul Scherrer Institute (PSI), 5232 Villigen-PSI, SwitzerlandShanghai Key Lab for Urban Ecological Processes and Eco-Restoration, School of Ecological and Environmental Sciences, East China Normal University, 200241 Shanghai, ChinaInstitute for Atmospheric and Earth System Research (INAR)/Physics, Faculty of Science, University of Helsinki, 00014 University of Helsinki, FinlandAerosol and Haze Laboratory, Beijing Advanced Innovation Center for Soft Matter Sciences and Engineering, Beijing University of Chemical Technology (BUCT), 100029 Beijing, ChinaJoint International Research Laboratory of Atmospheric and Earth System Sciences, School of Atmospheric Sciences, Nanjing University, 210023 Nanjing, ChinaInstitute for Atmospheric and Earth System Research (INAR)/Physics, Faculty of Science, University of Helsinki, 00014 University of Helsinki, FinlandDepartment of Computer Science, Faculty of Science, University of Helsinki, 00014 University of Helsinki, FinlandLaboratory of Atmospheric Chemistry, Paul Scherrer Institute (PSI), 5232 Villigen-PSI, Switzerland<p>The concentrations of atmospheric particulate matter and many of its constituents are temporally auto-correlated. However, this information has not been utilized in source apportionment methods. Here, we present a Bayesian matrix factorization model (BAMF) that considers the temporal auto-correlation of the components (sources) and provides a direct error estimation. The performance of BAMF is compared with positive matrix factorization (PMF) using synthetic Time-of-Flight Aerosol Chemical Speciation Monitor data, representing different urban environments from typical European towns to megacities. We find that BAMF resolves sources with overall higher factorization performance (temporal behavior and bias) than PMF on all datasets with temporally auto-correlated components. Highly correlated components continue to be challenging and ancillary information is still required to reach good factorizations. However, we demonstrate that adding even partial prior information about the chemical composition of the components to BAMF improves the factorization. Overall, BAMF-type models are promising tools for source apportionment and merit further research.</p>https://amt.copernicus.org/articles/17/1251/2024/amt-17-1251-2024.pdf |
| spellingShingle | A. Rusanen A. Rusanen A. Björklund M. I. Manousakas J. Jiang J. Jiang M. T. Kulmala M. T. Kulmala M. T. Kulmala K. Puolamäki K. Puolamäki K. R. Daellenbach A novel probabilistic source apportionment approach: Bayesian auto-correlated matrix factorization Atmospheric Measurement Techniques |
| title | A novel probabilistic source apportionment approach: Bayesian auto-correlated matrix factorization |
| title_full | A novel probabilistic source apportionment approach: Bayesian auto-correlated matrix factorization |
| title_fullStr | A novel probabilistic source apportionment approach: Bayesian auto-correlated matrix factorization |
| title_full_unstemmed | A novel probabilistic source apportionment approach: Bayesian auto-correlated matrix factorization |
| title_short | A novel probabilistic source apportionment approach: Bayesian auto-correlated matrix factorization |
| title_sort | novel probabilistic source apportionment approach bayesian auto correlated matrix factorization |
| url | https://amt.copernicus.org/articles/17/1251/2024/amt-17-1251-2024.pdf |
| work_keys_str_mv | AT arusanen anovelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT arusanen anovelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT abjorklund anovelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT mimanousakas anovelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT jjiang anovelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT jjiang anovelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT mtkulmala anovelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT mtkulmala anovelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT mtkulmala anovelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT kpuolamaki anovelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT kpuolamaki anovelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT krdaellenbach anovelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT arusanen novelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT arusanen novelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT abjorklund novelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT mimanousakas novelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT jjiang novelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT jjiang novelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT mtkulmala novelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT mtkulmala novelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT mtkulmala novelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT kpuolamaki novelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT kpuolamaki novelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization AT krdaellenbach novelprobabilisticsourceapportionmentapproachbayesianautocorrelatedmatrixfactorization |