You are here:
COMBINING DISPERSION MODELING AND BIG DATA FROM KANSAS CITY TRANSPORTATION AND LOCAL-SCALE AIR QUALITY STUDY (KC-TRAQS) FOR COMMUNITY-SCALE EXPOSURE CHARACTERIZATION
Citation:
Isakov, Vladilen, E. Kimbrough, S. Krabbe, M. Breen, S. Arunachalam, B. Naess, AND C. Seppanen. COMBINING DISPERSION MODELING AND BIG DATA FROM KANSAS CITY TRANSPORTATION AND LOCAL-SCALE AIR QUALITY STUDY (KC-TRAQS) FOR COMMUNITY-SCALE EXPOSURE CHARACTERIZATION. HARMO19 International Conference, Bruges, BELGIUM, June 03 - 06, 2019.
Impact/Purpose:
The Kansas City TRansportation and Local-scale Air Quality Study (KC-TRAQS) produced a rich dataset from both traditional measurement techniques and emerging measurement technologies. This big dataset with differing time and spatial resolutions requires innovative data analysis approaches to interpret. In addition, innovative data analysis approaches are needed for emission source attribution for near-source exposure assessments. In this study, we used dispersion modeling in support of the field study design and analysis. We are extending this work to use data fusion methods as follows to create accurate, fine-scale air quality characterization for the study domain: 1) fused field of observations including stationary, mobile measurements, and portable sensor technologies, and 2) fused field of observations from (1) and dispersion model estimates.
Description:
Spatially- and temporally-resolved air quality characterization is critical for community scale exposure studies and for developing future air quality mitigations. The Kansas City TRansportation local-scale Air Quality Study (KC-TRAQS), designed to characterize air pollution in a community affected by major transportation sources in Southeast Kansas City, Kansas, produced a rich dataset from both traditional measurement techniques and emerging measurement technologies. The study was conducted over a one-year period and included both traditional and lower-cost sampling methods, site-specific meteorological observations, and mobile measurements for multiple air pollutants (CO, CO2, NOx, PM2.5 and EC/OC components). The dataset includes more than two million rows of data which translates to approximately 108 million data values. This big dataset with differing time and spatial resolutions requires innovative data analysis approaches to interpret. In addition, innovative data analysis approaches are needed for emission source attribution for near-source exposure assessments. In this study, we used dispersion modeling in support of the field study design and analysis. In the study design phase, the presumptive placement of fixed monitoring sites and mobile monitoring routes have been corroborated using a research screening tool C-PORT, customized for KC-TRAQS application to assess the spatial and temporal coverage relative to the entire study area extent. In the analysis phase, dispersion modeling was used in combination with observations to help interpreting the KC-TRAQS data. We are extending this work to use data fusion methods as follows to create accurate, fine-scale air quality characterization for the study domain: 1) fused field of observations including stationary, mobile measurements, and portable sensor technologies, and 2) fused field of observations from (1) and dispersion model estimates. Using these two fused fields, we are developing methods to estimate relative source contribution of air pollution sources in the local community, and preliminary results will be presented
