Spatial and temporal distribution of sea salt aerosol mass concentrations in the marine boundary layer from the Arctic to the Antarctic

1. Sampling information The samples were collected from the Chinese Research Vessel (R/V) Xuelong, which undertook eight Antarctic cruises and five Arctic cruises, in summer, between 2008 and 2018. Briefly, all cruises used a large-flow air sampler to collect TSP (total suspended particulates) samples as well as field blanks. To avoid contamination, the large flow samplers were placed on the top deck during all cruises. All samples were collected using the relevant filter at a flow rate of 1.05 m3/min up to 57 h (most of them are 24h). After sampling, the samples were stored in a refrigerator at - 4 °C, prior to laboratory analysis. 2.Ion chromatography 467 samples were measured for ion concentrations. In summary, a punch sample of each filter was extracted by sonication and filtered through pre-washed 0.22-µm filters. Na+ in the samples was measured by ICS-2100 ion chromatography (Thermo Fisher Scientific Inc., USA). Each test had at least two reagent blanks, two field blanks, and one recovery. The recovery rate of all recovered samples was between 90% and 110%. The typical analytical precision of the instrument was better than 10 % relative standard deviation (RSD) for all ions. For comparison between the GEOS-Chem model and the observations, we converted the observed Na+ mass concentrations to SSA mass concentrations using a factor of 3.256, based on the mass ratio of Na+ in seawater. 3.GEOS-Chem model The GEOS-Chem global tropospheric chemistry model v12.2.6 (http://acmg.seas.harvard.edu/geos/) was used to simulate SSA concentrations with a spatial resolution of 2° × 2.5° at an hourly resolution (Text S3). Briefly, the model is driven by the Modern-Era Retrospective Analysis for Research and Applications-2 (MERRA-2). As our samples were TSP samples, the default SSA radius setting in the model (dp: particle diameter ~16 µm) was selected. SSA source functions in GEOS-Chem with an empirical dependence on SST derived by Jaegle et al. (2011). We simulated SSA concentrations every hour along the specific ship tracks and using the meteorology for the same dates as the measurements, and hourly-mean output for grid boxes extracted to match ship position. After simulating the hourly SSA concentration values along the ship tracks, all the simulated values during the sampling period (such as 24 hours) were arithmetic averaged to obtain the final SSA simulated value for each sample, which was compared with the observed value. 4.Auxiliary data The U10 data analyzed in this study were obtained with the meteorological detection system atop R/V Xuelong (33.5 m) during the Arctic and Antarctic cruises. SST, T2M (temperature at 2 m in the atmosphere), relative humidity (RH), sea ice fraction (FRESEAICE) were also acquired from the MERRA-2 meteorological field in the GEOS-Chem model, more details about the MERRA-2 data were given in Molod et al. (2015)