A new metric space incorporating radon-222 for generation of back trajectory clusters in atmospheric pollution studies

A novel metric space for the clustering of back trajectories to be used in fine particle aerosol data analysis was proposed and evaluated. The metric is based on spatial and non-spatial variables incorporating great-circle distance, altitude and radon-222.Its performance was examined using the intra-cluster variation of measured and fingerprint apportioned aerosol mass as the quantitative criterion. The new metric was demonstrated to perform better than those based on great-circle distance, or a great-circle distance and altitude alone. The same criterion was applied to investigate the clustering performance as a function of the length of its back trajectories. The optimum back trajectory length was found to be dependent on the pollution source being considered.Performance tests, as well as the application of the new metric space to re-analysis of previously published results, were based on a three year long dataset comprising co-located aerosol fine particles (PM2.5) collection and hourly measurements of radon-222 concentration.The new metric space can easily be redefined to include other trace species.