Annual Mean PM_2.5 Trace Elements 50m Grids in Urban Areas and 1km Grids in Non-Urban Areas for Contiguous U.S., 2000-2019, v1

Particulate matter with aerodynamic diameter less than 2.5 µm (PM_2.5) is a multi-million human silent killer worldwide, and contains many trace elements (TEs). Understanding the relative toxicity is largely limited by lack of data. Ensembles of machine learning models were used to generate ~163 billion predictions estimating annual mean PM_2.5 TEs, namely bromine (Br), calcium (Ca), copper (Cu), iron (Fe), potassium (K), nickel (Ni), lead (Pb), silicon (Si), vanadium (V), and zinc (Zn) across 3,535 contiguous US urban areas at 50 m spatial resolution and at 1 km in non-urban areas from 2000-2019. The results highlight substantial intra-urban and inter-urban variations, shrinkages, stagnations, or expansions of hotspots, and trends across 20 years. This data opens avenues for future research from epidemiological studies to environmental justice analyses and more.

The monitored data from ~600 locations were integrated with 160+ predictors, such as time and location, satellite observations, composite predictors, meteorological covariates, and many novel land use variables using several machine learning algorithms and ensemble methods. The monitoring data were divided into training (70%) and test (30%) sets. Multiple machine-learning models (random forest (RF), stochastic gradient boosting (GBM), extreme gradient boosting (XGB), cubist, or K-nearest neighbors (KNN)) were developed covering 3,535 urban areas where the majority (~80%) of US population lives at 50 m spatial resolution and at 1 km in non-urban areas, and their predictions were esembled using either a generalized additive model (GAM) ensemble geographically-weighted-averaging (GAM-ENWA) or super-learners (SLs; RF-SL, GBM-SL, XGB-SL, cubist-SL, KNN-SL, or support vector machines (SVM-SL)). The overall best model R² values on the test sets ranged from 0.79 for Cu using RF-SL to 0.88 for Zn using GAM-ENWA in non-urban areas. These in urban models ranged from 0.80 for Cu using SVM-SL to 0.88 for Zn using GAM-ENWA. The Coordinate Reference System (CRS) for predictions is World Geodetic System 1984 (WGS84) https://epsg.io/4326. The units for TE predictions are nanograms per cubic meter (ng/m3).

Download the Annual Mean PM_2.5 Trace Elements 50m Grids in Urban Areas and 1km Grids in Non-Urban Areas for Contiguous U.S., 2000-2019, v1. The files in the table below are in RDS format. The urban file sizes range from approximately 3.6 to 3.8 GB and the non-urban file sizes are about 145 MB.

Amini, H.^{1, 2*}, M. Danesh-Yazdi^1,3, Q. Di⁴, W. Requia⁵, Y. Wei¹, Y. AbuAwad⁶, L. Shi⁷, M. Franklin⁸, C.-M. Kang¹, J. M. Wolfson¹, P. James^9,1, R. Habre¹⁰, Q. Zhu⁷, J. S. Apte^11,12, Z. J. Andersen², X. Xing¹³, C. Hultquist^13,14, I. Kloog¹⁵, F. Dominici^1,16, P. Koutrakis¹, J. Schwartz¹. 2022. Annual Mean PM_2.5 Trace Elements 50m Grids in Urban Areas and 1km Grids in Non-Urban Areas for Contiguous U.S., 2000-2019, v1. (Preliminary Release). Palisades, NY: NASA Socioeconomic Data and Applications Center (SEDAC). https://doi.org/10.7927/1x94-mv38. Accessed DAY MONTH YEAR.

¹Harvard T.H. Chan School of Public Health, Boston, MA, United States
²Department of Public Health, University of Copenhagen, Copenhagen, Denmark
³Stony Brook University, New York, USA
⁴Vanke School of Public Health, Tsinghua University, Beijing, China
⁵Fundação Getúlio Vargas, Brasilia, Brazil
⁶PERFORM Centre, Concordia University, Montreal, Quebec, Canada
⁷Gangarosa Department of Environmental Health, Rollins School of Public Health, Emory University, Atlanta, GA, United States
⁸Department of Statistical Sciences, University of Toronto, Toronto, Canada
⁹Harvard Medical School and Harvard Pilgrim Health Care Institute, Boston, MA, United States
¹⁰Keck School of Medicine, University of Southern California, Los Angeles, CA, United States
¹¹Department of Civil and Environmental Engineering, University of California, Berkeley, CA, United States
¹²School of Public Health, University of California, Berkeley, CA, United States
¹³Center for International Earth Science Information Network (CIESIN), Columbia University, Palisades, NY, United States
¹⁴School of Earth and Environment, University of Canterbury, Christchurch, New Zealand
¹⁵Icahn School of Medicine at Mount Sinai, NY, United States
¹⁶Harvard Data Science Initiative, Cambridge, MA, United States

This project was supported by Cyprus Harvard Endowment Program for the Environment and Public Health,Novo Nordisk Foundation Challenge Programme grant NNF17OC0027812, U.S. EPA grant RD-8358720, NIH grants R01AG074357, R01 HL150119, R01MD012769, R01 ES028033, 1R01AG060232-01A1, 1R01ES030616, 1R01AG066793-01R01, and R01ES028033-S1, Fernholz Foundation, U.S. EPA grant RD-835872, NIH grants P30 ES000002 and R01ES032418-01. The contents are solely the responsibility of the grantees and do not necessarily represent the official views of the U.S. EPA. Further, U.S. EPA does not endorse the purchase of any commercial products or services mentioned in the data or documents.

This is a preliminary open data release, pending peer review of the associated journal article. Following the peer review process, data curation will be completed by the NASA Socioeconomic Data and Applications Center (SEDAC) and the data will be disseminated through the SEDAC catalog.