Baseline characterisation of source contributions to daily-integrated PM_2.5 observations at Cape Grim using Radon-222

We discuss 15 years (2000-2015) of daily-integrated PM_2.5 samples from the Cape Grim Station. Ion beam analysis and positive matrix factorisation are used to identify six source-type fingerprints: fresh sea salt (57%); secondary sulfate (14%); smoke (13%); aged sea salt (12%); soil dust (2.4%); and industrial metals (1.5%). An existing hourly radon-only baseline selection technique is modified for use with the daily-integrated observations. Results were not significantly different for days on which >20 hours were below the baseline radon threshold compared with days when all 24 hours satisfied the baseline criteria. This relaxed daily baseline criteria increased the number of samples for analysis by almost a factor of two. Two radon baseline thresholds were tested: historic (100 mBq m^-3), and revised (50 mBq m^-3). Median aerosol concentrations were similar for both radon thresholds, but maximum values were higher for the 100 mBq m^-3 threshold. Back trajectories indicated more interaction with southern Australia and the Antarctic coastline for air masses selected with the 100 mBq m^-3 threshold. Radon-only baseline selection using the 50 mBq m^-3 threshold was more selective of minimal terrestrial influence than a similar recent study using wind direction and back trajectories. The ratio of concentrations between terrestrial and baseline days for the primary sources soil, smoke and industrial metals was 3.4, 2.6, and 5.5, respectively. Seasonal cycles of soil dust had a summer maximum and winter minimum. Seasonal cycles of smoke were of similar amplitude for terrestrial and baseline events, but of completely different shape: peaking in autumn and spring for terrestrial events, compared to summer for baseline conditions. Seasonal cycles of industrial metals had a summer maximum and winter minimum. A significant fraction of the Cape Grim baseline smoke and industrial metal contributions appeared to be derived from long-term transport (>3 weeks since last terrestrial influence).