Regional polycentric spatial patterns and air pollution: An empirical study of German metropolitan regions
Published in Cities, 2026 · Download PDF

This study empirically examines the effectiveness of polycentric urban spatial patterns in reducing air pollution across German metropolitan regions, using satellite-based air quality measures, multi-year datasets and a multi-spatial scale research design. Our findings suggest that polycentric development does not significantly impact regional air pollution overall. However, an intra-regional analysis reveals that urban cores in more polycentric regions experience higher pollution levels compared to their monocentric counterparts. When disaggregating by regional size, we find that polycentricity has no influence on air quality in smaller regions. However, this relationship does not hold as regional size increases, with additional polycentric development leading to higher pollution levels in medium and large regions, particularly in urban core areas. Although we find that differences in population density can explain pollution in the peripheries of polycentric regions, this explanation does not extend to urban core areas. In this case, we find that differences in air quality may be due to the cumulative effects of congestion in larger urban cores, where strong functional links among multiple centers intensify commuting demands and traffic congestion more than in peripheral areas. This prolonged and more congested travel likely contributes to higher pollution levels in urban core areas of polycentric regions. In contrast, smaller regions may avoid these negative externalities because their scale of commuting is insufficient to generate the extended travel times and congestion.
Keywords: polycentricity; PM2.5 concentrations; urban spatial structure; commuting flow patterns; population density
Recommended citation: Li, W., Schmidt, S., & Gu, J. (2026). Regional polycentric spatial patterns and air pollution: An empirical study of German metropolitan regions. Cities, 170, 106676.
