A comparison of statistical and machine learning methods for creating national daily maps of ambient PM2.5 concentration

Veronica J. Berrocal, Yawen Guan, Amanda Muyskens, Haoyu Wang, Brian J. Reich, James A. Mulholland, Howard H. Chang

Research output: Contribution to journalArticlepeer-review

47 Scopus citations

Abstract

A typical challenge in air pollution epidemiology is to perform detailed exposure assessment for individuals for which health data are available. To address this problem, in the last few years, substantial research efforts have been placed in developing statistical methods or machine learning techniques to generate estimates of air pollution at fine spatial and temporal scales (daily, usually) with complete coverage. However, it is not clear how much the predicted exposures yielded by the various methods differ, and which method generates more reliable estimates. In this paper, we aim to address this gap by evaluating a variety of exposure modeling approaches, comparing their predictive performance. Using PM2.5 in year 2011 over the continental U.S. as a case study, we generate national maps of ambient PM2.5 concentration using: (i) ordinary least squares and inverse distance weighting; (ii) kriging; (iii) statistical downscaling models, that is, spatial statistical models that use the information contained in air quality model outputs; (iv) land use regression, that is, linear regression modeling approaches that leverage the information in Geographical Information System (GIS) covariates; and (v) machine learning methods, such as neural networks, random forests and support vector regression. We examine the various methods’ predictive performance via cross-validation using Root Mean Squared Error, Mean Absolute Deviation, Pearson correlation, and Mean Spatial Pearson Correlation. Additionally, we evaluated whether factors such as, season, urbanicity, and levels of PM2.5 concentration (low, medium or high) affected the performance of the different methods. Overall, statistical methods that explicitly modeled the spatial correlation, e.g. universal kriging and the downscaler model, outperform all the other exposure assessment approaches regardless of season, urbanicity and PM2.5 concentration level. We posit that the better predictive performance of spatial statistical models over machine learning methods is due to the fact that they explicitly account for spatial dependence, thus borrowing information from neighboring observations. In light of our findings, we suggest that future exposure assessment methods for regional PM2.5 incorporate information from neighboring sites when deriving predictions at unsampled locations or attempt to account for spatial dependence.
Original languageEnglish (US)
Pages (from-to)117130
JournalAtmospheric Environment
Volume222
DOIs
StatePublished - Feb 2020
Externally publishedYes

Bibliographical note

KAUST Repository Item: Exported on 2021-02-08
Acknowledgements: The authors thank Drs. Xuefie Hu and Yang Liu for providing the land use and meteorology data, and Niru Senthilkumar for providing the CMAQ and air quality data. This material was based upon work partially supported by the National Science Foundation under Grant DMS-1638521 to the Statistical and Applied Mathematical Sciences Institute. BJR is also supported by NIH R01ES027892, DOI-JFSP 14-1-04-9 and KAUST 3800.2; VJB is supported by NIH P30ES017885. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Institutes of Health and the National Science Foundation, U.S.
This publication acknowledges KAUST support, but has no KAUST affiliated authors.

Fingerprint

Dive into the research topics of 'A comparison of statistical and machine learning methods for creating national daily maps of ambient PM2.5 concentration'. Together they form a unique fingerprint.

Cite this