Disaggregating gridded air quality data for individual exposure modelling

Gerharz Lydia, Gräler Benedikt, Pebesma Edzer

Abstract

This paper presents an analysis of disaggregation for PM10 pollution from a grid to point support for exposure modelling on a GPS track representing an individual space-time trajectory. Different sets of explanatory variables were tested to predict spatial variability of mobile PM10 measurements at the point support. Disaggregation was performed using unconditional Gaussian simulation. The results show a considerable amount of uncertainty added due to disaggregation that depends in strength on the auxiliary data set used for prediction. Subsequent aggregation over the GPS track leads to a reduction in uncertainties.