Aerosol sample collection: Aerosols were collected on cellulose filters using a Tisch Series 235 high-volume (~1 m3 air min-1) aerosol sampler equipped with a cascade impactor designed for the separation of coarse (>1 µm) and fine (<1 µm) aerosol fractions. The sampler was mounted on a platform atop the ship’s wheelhouse as far forward as possible. Relative wind speed and direction were monitored during aerosol sample collection, and the sampler was operated only when the ship was steaming into the prevailing wind, in an effort to avoid contamination from the ship’s exhaust and superstructure. The cascade impactor was loaded with Whatman 41 cellulose filters that had been pre-cleaned at Old Dominion University using 0.1 N and 0.5 N hydrochloric acid, following a procedure modified from Baker et al. [2006]. Six filters (five 25 cm x 25 cm slotted sheets and one 20.3 cm x 25.4 cm backing sheet) were used for the collection of each aerosol sample. Air flow rates were calculated using manufacturer-provided flow conversion tables and the pressure decrease across the filters, which was measured at the start and end of sample collection using a handheld digital manometer (Dwyer Series). The total air volume sampled was estimated from the period of sample collection and the average value of the initial and final flow rates.

Aerosol sample processing: Immediately following collection, the aerosol-laden filters were unloaded and subsampled within a Class-100 clean air bench. Fixed portions of the aerosol-laden cellulose filters corresponding to each aerosol size fraction (<1 µm and >1 µm) were transferred into 47 mm diameter perfluoroalkoxy alkane filter funnel assemblies (Savillex) loaded with acid-cleaned 0.4 µm polycarbonate filter membranes, and leached with 750 ml of 18.2 MΩ-cm resistivity deionized water (Barnstead Nanopure), in a flow-through protocol modfied after Buck et al. [2006]. Aliquots of the leachate solutions were immediately transferred into (i) acid-cleaned 125 ml low-density polyethylene bottles (Nalgene), then acidified with 500 μl 6N ultrapure hydrochloric acid (Fisher Chemical, Optima), for post-cruise analysis of soluble aerosol iron, and (ii) 60 mL polypropylene tubes (Falcon) for shipboard analysis of soluble aerosol nitrate+nitrite, phosphate and ammonium. In addition, separate portions of the aerosol-laden cellulose filters corresponding to each aerosol size fraction (<1 µm and >1 µm) were stored in pre-cleaned ziploc polyethylene bags for microwave acid digestion at Old Dominion University. The microwave acid digestion procedure was adapted from Morton et al. [2013], with the resulting digest solutions evaporated to near dryness and then diluted with 20 mL of 1% ultrapure nitric acid (Fisher Chemical, Optima).

Rainwater sample collection: Two methods were used to collect the rain samples at sea. Samples Rain-01 and Rain-02 were collected in two acid-cleaned 2 L wide-mouth fluorinated high-density polyethylene bottles (Nalgene) mounted inside a polyethylene bucket, using an N-Con Systems automated rain sampler. Samples Rain-03, Rain-04 and Rain-05 were manually collected using an acid-cleaned high-density polyethylene funnel (Nalgene) connected by a Teflon collar to an acid-cleaned 2 L low-density polyethylene bottle (Nalgene). Both sample collectors were mounted on a platform atop the ship’s wheelhouse as far forward as possible, with samples collected whilst the ship was steaming into the prevailing wind, in an effort to avoid contamination from the ship’s exhaust and superstructure. The ODU Rain Composite combines samples collected on the Old Dominion University campus during summer 2014 using the manual funnel sampling method.

Rainwater sample processing: Immediately following collection, rainwater sample containers were capped and transferred to a shipboard Class-100 clean air bench for processing. From each sample, aliquots were transferred into (i) 60 mL polypropylene tubes (Falcon), which were frozen for post cruise analysis of nitrate+nitrite, phosphate and ammonium, and, when there was sufficient sample volume (ii) acid-cleaned 125 ml low-density polyethylene bottles (Nalgene) and acidified to pH 1.8 with 6N ultrapure hydrochloric acid (Fisher Chemical, Optima) for post-cruise analysis of total iron (more strictly, total acid-labile iron). In addition, if the volume of sample was sufficient, the rainwater was filtered through an acid-cleaned 0.4 µm polycarbonate membrane using a 47 mm diameter perfluoroalkoxy alkane filter funnel assembly (Savillex), and the filtrate transferred into an acid-cleaned 125 ml low-density polyethylene bottles (Nalgene) and acidified to pH 1.8 with 6N ultrapure hydrochloric acid (Fisher Chemical, Optima) for post-cruise analysis of dissolved iron.

DFe and TFe: Dissolved iron and total iron was determined in aerosol leachate solutions (DFe), aerosol digest solutions (TFe), filtered rainwater (DFe) and unfiltered rainwater (TFe, or more strictly, total acid-labile iron) were determined at Old Dominion University using a ThermoFisher Element2 high-resolution inductively-coupled plasma mass spectrometer (HR-ICP-MS). Sample solutions were introduced into the ICP-MS without preconcentration, and quantified using matrix-matched external standard solutions prepared with SPEX CertiPrep Claritas PPT grade standards. An indium internal standard was used to correct for instrumental drift. Analytical limits of detection are estimated as 0.57 nM for DFe and for TFe in ulfiltered rainwater, and 53 nM for TFe in the aerosol digest solutions. The values presented for DFe and TFe in samples represent concentrations after subtracting the concentrations of the corresponding field blanks (Rain-Sampler-Blank, Rain-Funnel-Blank, or Aer-Blank). Atmospheric loadings of soluble aerosol iron (Sol Aer Fe) and total aerosol iron (Tot Aer Fe) were calculated from DFe in aerosol leachates and TFe in aerosol digest solutions, respectively, and the total air volume sampled in each case, after correcting for the fraction of the active filter area that was leached (for DFe) or digested (for TFe).

NO3+NO2: Dissolved nitrate plus nitrite was determined in aerosol leachate solutions and rainwater with an Astoria Pacific nutrient autoanalyzer, using standard colorimetric methods with an estimated detection limit of 0.14 µM (Parsons et al., 1984; Price and Harrison, 1987). Atmospheric loadings of soluble aerosol nitrate plus nitrite (Sol Aer NO3+NO2) were calculated from NO3+NO2 in aerosol leachates and the total air volume sampled, after correcting for the fraction of the active filter area that was leached.

PO4: Dissolved phosphate was determined in aerosol leachate solutions and rainwater with an Astoria Pacific nutrient autoanalyzer, using standard colorimetric methods with an estimated detection limit of 0.03 µM (Parsons et al., 1984; Price and Harrison, 1987). Atmospheric loadings of soluble aerosol phosphate (Sol Aer PO4) were calculated from PO4 in aerosol leachates and the total air volume sampled, after correcting for the fraction of the active filter area that was leached.

NH4: Dissolved ammonium was determined in aerosol leachate solutions and rainwater using the manual orthophthaldialdehyde method (Holmes et al., 1999), with an estimated detection limit of 10 nM. Atmospheric loadings of soluble aerosol ammonium (Sol Aer NH4) were calculated from NH4 in aerosol leachates and the total air volume sampled, after correcting for the fraction of the active filter area that was leached.