Biases in long-term NO2 averages inferred from satellite observations due to cloud selection criteria

REMOTE SENSING OF ENVIRONMENT

卷:124

Article

Geddes, Jeffrey A.¹  Murphy, Jennifer G.¹  O'Brien, Jason M.²  Celarier, Edward A.³ 

[1] Univ Toronto, Dept Chem, Toronto, ON M5S 1A1, Canada

[2] Environm Canada, Air Qual Res Div, Toronto, ON, Canada

[3] NASA, Goddard Space Flight Ctr, Univ Space Res Assoc, Greenbelt, MD 20771 USA

关键词: Nitrogen dioxide, NO2;    Ground level pollution;    Satellite;    Ozone Monitoring Instrument, OMI;    Cloud fraction;    Remote sensing;    Cloud bias;    Toronto;    Air quality;    Vertical column density;   

DOI  :  10.1016/j.rse.2012.05.008

来源: Elsevier

【 摘 要 】

Retrievals of atmospheric trace gas column densities from space are compromised by the presence of clouds, requiring most studies to exclude observations with significant cloud fractions in the instrument's field of view. Using NO2 observations at three ground stations representing urban, suburban, and rural environments, and tropospheric vertical column densities measured by the Ozone Monitoring Instrument (OMI) over each site, we show that the observations from space represent monthly averaged ground-level pollutant conditions well (R = 0.86) under relatively cloud-free conditions. However, by analyzing the ground-level data and applying the OMI cloud fraction as a filter, we show there is a significant bias in long-term averaged NO2 as a result of removing the data during cloudy conditions. For the ground-based sites considered in this study, excluding observations on days when OMI-derived cloud fractions were greater than 0.2 causes 12:00-14:00 mean summer mixing ratios to be underestimated by 12% +/- 6%, 20% +/- 7%, and 40% +/- 10% on average (+/- 1 standard deviation) at the urban, suburban, and rural sites respectively. This bias was investigated in particular at the rural site, a region where pollutant transport is the main source of NO2, and where long-term observations of NOy were also available. Evidence of changing photochemical conditions and a correlation between clear skies and the transport of cleaner air masses play key roles in explaining the bias. The magnitude of a bias is expected to vary from site to site depending on meteorology and proximity to NOx sources, and decreases when longer averaging times of ground station data (e.g. 24-h) are used for the comparison. (c) 2012 Elsevier Inc. All rights reserved.

【 授权许可】

Free