Potential application of VIIRS Day/Night Band for monitoring nighttime surface PM_2.5 air quality from space

Authors

Jun Wang, University of Nebraska-Lincoln
Clint Aegerter, University of Nebraska-LincolnFollow
Xiaoguang Xu, University of Nebraska-LincolnFollow
James J. Szykman, U.S. Environmental Protection Agency

Document Type

Article

Date of this Version

2016

Citation

Atmospheric Environment 124 (2016), pp. 55-63, http://dx.doi.org/10.1016/j.atmosenv.2015.11.013.

Comments

U.S. government work.

Abstract

A pilot study is conducted to illustrate the potential of using radiance data collected by the Day/Night Band (DNB) of the Visible Infrared Imaging Radiometer Suite (VIIRS) aboard the Suomi National Polarorbiting Partnership (S-NPP) satellite for particulate matter (PM) air quality monitoring at night. The study focuses on the moonless and cloudless nights in Atlanta, Georgia during August-October 2012.We show with radiative transfer calculations that DNB at night is sensitive to the change of aerosols and much less sensitive to the change of water vapor in the atmosphere illuminated by common outdoor light bulbs at the surface. We further show both qualitatively that the contrast of DNB images can indicate the change of air quality at the urban scale, and quantitatively that change of light intensity during the night (as characterized by VIIRS DNB) reflects the change of surface PM_2.5. Compared to four meteorological variables (u and v components of surface wind speed, surface pressure, and columnar water vapor amount) that can be obtained from surface measurements, the DNB light intensity is the only variable that shows either the largest or second largest correlation with surface PM_2.5 measured at 5 different sites. A simple multivariate regression model with consideration of the change of DNB light intensity can yield improved estimate of surface PM_2.5 as compared to the model with consideration of meteorological variables only. Cross validation of this DNB-based regression model shows that the estimated surface PM_2.5 concentration has nearly no bias and a linear correlation coefficient (R) of 0.67 with respect to the corresponding hourly observed surface PM_2.5 concentration. Furthermore, groundbased observations support that surface PM_2.5 concentration at the VIIRS night overpass (~1:00 am local) time is representative of daily-mean PM_2.5 air quality (R = 0.82 and mean bias of -0.1 mg m^-3). While the potential appears promising, mapping surface PM_2.5 from space with visible light at night still face various challenges and the strategies to address some of these challenges are elaborated for future studies.