ETC/ACM Report 18/2018: Land cover and traffic data inclusion in PM mapping

25 Jun 2019

Jan Horálek, Peter de Smet, Philipp Schneider, Frank de Leeuw, Marketa Schreiberova

Prepared by: ETC/ACM members Jan Horálek1, Peter de Smet2, Philipp Schneider3, Frank de Leeuw2, Markéta Schreiberová1

1Czech Hydrometeorological Institute (CHMI), Prague, Czech Republic; 2National Institute for Public Health and the Environment (RIVM), Bilthoven, Netherlands; 3Norwegian Institute of Air Research (NILU), Norway

Annual European-wide air quality maps have been produced using geostatistical techniques for many years and is based primarily on air quality measurements. The mapping method follows in principle the sequence of regression – interpolation – merging. It combines monitoring data, chemical transport model outputs and other supplementary data (such as altitude and meteorology) using a linear regression model followed by kriging of its residuals (‘residual kriging’), applied separately for rural and urban background areas. The rural and urban background map layers are
subsequently merged on basis of population densities into one final concentration map for Europe.

Inclusion of land cover and road type data among the set of the supplementary data demonstrated to improve the quality of urban and rural background layers in the NO2 map and is currently routinely applied in the NO2 mapping. In addition, an urban traffic map layer based on the measurement data from traffic stations is constructed and takes art in the merging process with the rural and urban background map layers to reach a final NO2 map.

This report examines now – due to its proved added value in the NO2 mapping – whether for PM10 and PM2.5 the similar method provides also sufficient added value to include it on a routinely basis in the production of the final concentration map and population exposure estimates.. It concerns the inclusion of land cover data and road type data in the background map layers, as well as the inclusion of the urban traffic layer based on traffic measurement stations. The analysis is done based on 2015 data, being the most recent year with all data needed available when this study started.