Urban emissions hotspots: Quantifying vehicle congestion and air pollution using mobile phone GPS data.
Data collected during the 2011 DISCOVER-AQ field campaign in the Baltimore Washington region were used to evaluate CO and NOx emissions in the National Emissions Inventory (NEI). The average emissions ratio for the region was seen to be 11.2 ± 1.2 mol CO/mol NOx, 21% higher than that predicted by the NEI. Comparisons between in situ and remote observations and CMAQ model output show agreement in CO emissions of 15 ± 11% while NOx emissions are overestimated by 51e70% in Maryland. Satellite observations of CO by MOPITT show agreement with the Community Multiscale Air Quality (CMAQ) model within 3% over most of the eastern United States. CMAQ NOy mixing ratios were a factor of two higher than observations and result from a combination of errors in emissions and PAN and alkyl nitrate chemistry, as shown by comparison of three CMAQ model runs. Point source NOx emissions are monitored and agree with modeled emissions within 1% on a monthly basis. Because of this accuracy and the NEI assertion that approximately 3/4 of emissions in the Baltimore Washington region are from mobile sources, the MOVES model's treatment of emissions from aging vehicles should be investigated; the NEI overestimate of NOx emissions could indicate that engines produce less NOx and catalytic converters degrade more slowly than assumed by MOVES2010. The recently released 2011 NEI has an even lower CO/NOx emissions ratio than the projection used in this study; it overestimates NOx emissions by an even larger margin. The implications of these findings for US air quality policy are that NOx concentrations near areas of heavy traffic are overestimated and ozone production rates in these locations are slower than models indicate. Results also indicate that ambient ozone concentrations will respond more efficiently to NOx emissions controls but additional sources may need to be targeted for reductions. © 2014 Elsevier Ltd. All rights reserved.