Temporal and spatial variations in dust activity in Australia based on remote sensing and reanalysis data sets

Abstract. Spatial and temporal variations in the level of dust activity can provide valuable information for policy making and climate research. Recently, MODIS aerosol products have been successfully used for retrieving dust aerosol optical depth (DAOD), especially over bright dust source areas and MERRA-2 aerosol reanalysis provides DAOD, and additionally other dust aerosol-related parameters. In this study, spatial and temporal variations in dust activity in Australia were analyzed using MODIS and MERRA-2 combined (M&M) DAOD and MERRA-2 near-surface dust concentrations/estimated PM10 for the period from 1980–2020. Validation results show that M&M DAOD has an expected error of ±(0.016+0.15τ) compared to the ground observations at the AERONET sites. MERRA-2 near-surface dust concentrations show a power law relationship with visibility data collected at meteorological stations with an r² value from 0.18 to 0.44, and the estimated MERRA-2 PM10 shows similar temporal variations and correlates with ground-based PM10 data with an r² value from 0.14 to 0.44 at six selected stations in Australia. Moreover, MERRA-2 dust flux shows the same major dust pathways as those in previous studies and similar dust emissions/deposition areas. Dust events based on DAOD over eastern Australia are concentrated in the north in December, in the south in February, and can occur anywhere in January. Near-surface dust concentration was found to be the highest (over 200 μg/m³) over the center of Lake Eyre Basin in central Australia and radially decreased to the coast to below 20 μg/m³ via the two main pathways in the southwest and northeast. The ratio of near-surface dust concentration to PM10 shows a similar spatial pattern. Total dust emission was estimated to be 40 MT (mega-tonnes) per year over the period 1980–2020, of which nearly 50 % was deposited on land and the rest exported away from the Australian continent.