Environmental air quality impacts human well-being and disease, but the availability of air quality data is limited to selected locations because of the complexity involved in its measurement. Among air pollutants, PM2.5 is of the greatest concern. These particles are not captured by the lungs'natural defenses and can be inhaled deeply, where they can cause health problems ranging from asthma attacks to heart disease. PM2.5 currently is measured primarily by ground monitoring stations located at approximately 320 EPA sites, providing limited local geographic coverage. However, there are satellites that make a variety of aerosol observations and provide a daily global picture of atmospheric particulates in the form of aerosol optical depth (AOD). Scientists have attempted to compute ground-level PM2.5 (GLP) from these AOD data. However, the multivariate nonlinear relationship between AOD and PM2.5 imposes limitations in computing GLP using satellite data. This project proposes to overcome these limitations by computing reliable GLP via a new methodology which has already been tested and validated. The study has two specific aims: (1) develop satellite-derived daily GLP estimates for the contiguous U.S., and (2) examine spatial and temporal associations between GLP exposure and hospital visits for asthma exacerbation in Mississippi. Using our methodology, we will generate daily GLP data for a 12-month period at a resolution of 0.10x0.10 (~10x10km2), providing approximately 82,000 data points as opposed to about 300 data points available daily from EPA ground monitoring stations within the contiguous U.S. We will address the nonlinear relationship between PM2.5 and AOD which is a function of humidity, temperature, surface pressure, surface wind speed, surface type, boundary layer height, and AOD by accounting for these variables using a machine learning process. The AOD that will be used in this process will be generated by merging AOD data from multiple satellite sensors. Meteorological data will come from NOAA NCEP. Surface type data will be obtained from the satellite-identified vegetation index. Boundary layer height, which is the mixed layer of the atmosphere closest to the ground where people live and work, will come from CALIPSO data which provides vertical profiles of atmospheric aerosol extinction. Information on GLP levels will allow the scientific community to better understand health impacts from exposure to low, moderate, or high levels of PM2.5. Moreover, in places where PM2.5 levels are elevated only occasionally, such as Mississippi, the short-term health impact of increases in PM2.5 can be studied more precisely. National GLP data will be made available to other researchers to facilitate future explorations of how PM2.5 exposure impacts a wide range of health conditions, thereby making possible more timely prophylactic treatment, improving healthcare system preparedness, and better informing public health policymaking.

Public Health Relevance

Among air pollutants, particulates 2.5 micrometers in diameter and smaller (PM2.5) have the greatest impact on human health because they are small enough to be inhaled deeply into the lungs, making allergies, asthma, and other respiratory conditions worse. The proposed study will provide daily estimates of ground-level PM2.5 for the entire contiguous U.S. so that researchers can study how it affects disease at various locations in the country. The knowledge that results can be used to better inform people about the links between air pollution and disease, and to guide the development of air quality standards to improve public health.

National Institute of Health (NIH)
National Institute of Environmental Health Sciences (NIEHS)
Exploratory/Developmental Grants (R21)
Project #
Application #
Study Section
Infectious Diseases, Reproductive Health, Asthma and Pulmonary Conditions Study Section (IRAP)
Program Officer
Dilworth, Caroline H
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
University of Mississippi Medical Center
Other Health Professions
Schools of Allied Health Profes
United States
Zip Code
Lary, D J; Lary, T; Sattler, B (2015) Using Machine Learning to Estimate Global PM2.5 for Environmental Health Studies. Environ Health Insights 9:41-52
Lary, David J; Faruque, Fazlay S; Malakar, Nabin et al. (2014) Estimating the global abundance of ground level presence of particulate matter (PM2.5). Geospat Health 8:S611-30