Daily estimation of NO2 concentrations using digital tachograph data.

Abstract

Traffic information is crucial for estimating NO2 concentrations, but it is static and limited in predicting constantly changing NO2 levels. To overcome these challenges, this study utilized real-time spatial big data to capture both the spatial and temporal fluctuations in traffic. Digital tachograph (DTG) data, sourced from digital devices in all commercial vehicles, are employed to construct a DTG land use regression (LUR) model, and its performance is compared with that of a non-DTG-LUR model. The DTG-LUR model exhibits superior performance, with an explanatory power of 0.46, in contrast to the 0.36 of the non-DTG model. This significant improvement stems from the spatially and temporally dynamic DTG variables such as cargo traffic. This study introduces a novel approach for incorporating DTG data in correlating with NO2 concentrations. It underscores the advantage of DTG data in predicting daily NO2 fluctuations at a precise 200-m grid, which is not feasible with conventional data. The findings of the study highlight the immense potential of spatial big data for fine-grained analyses, which could enable hourly predictions of air pollution.