Sampling and Analytical Methodology:
Nutrient and aerosol addition bioassay experiments were carried out over 3 days in February 2012. Seawater was collected from offshore (water depth >700 m) outside the Bellairs Research Institute at West Barbados (13^o11.309’N, 59^o38.267’W). Surface water was pumped into acid cleaned sample rinsed carboys using a peristaltic pump with acid washed Teflon tubing and pre filtered through a 50 um mesh acid washed Nitex^© net to remove grazers. The seawater was stored in the dark until transport to the lab (within <2 hours). Seawater was dispensed into acid washed and sample rinsed polycarbonate bottles (500 mL each), pre-labeled with treatment type (12-20 bottles per treatment). Treatments included single nutrient (N, P, Fe) additions as well as a combination of N and P and a combination of N and Fe at concentrations representative of deep water in this area. Three aerosol treatments were used in this study representing aerosols deposited in three seasons, winter, spring and summer. Aerosols representing each of the seasons were added at concentrations simulating high and low deposition rates. High deposition was calculated to represent the cumulative deposition flux over 10 days of a strong dust storm event over the North Atlantic (300 g m^-2 yr^-1) to the upper 10 m mixed layer. Low deposition treatments were equivalent to the normal average deposition rate for Barbados (10 g m^-2 yr^-1) during spring and summer. A control (no addition, blank filter) treatment and procedural blanks (Milli-Q water) were also included. All bottles were incubated in a pool filled with circulating seawater to maintain local surface ocean temperature. The pool was covered with a neutral density shading screen to reduce light intensity by 50%. Water samples used for the experiment (pre additions) was collected to characterize the baseline conditions (baseline, 5 replicates) and 3 replicate bottles for each treatment were also collected immediately after the additions were administered (time zero, t0). The experiment took place over 3 days, and each day 3 (for nutrients) or 5 (for aerosols) randomly selected bottles for each treatment were collected at 4pm in the afternoon (e.g. time points t1-t3). Immediately upon collection each bottle was sampled for chlorophyll a, flow cytometry, nutrients, and trace metal concentrations.

60 mL subsamples of the 0.2 um filtered water were collected in acid washed sample rinsed LDPE bottles for trace metal analyses. Seawater samples were acidified to pH<2.0 by adding 45 uL concentrated trace metal grad HNO₃at least24 hours prior to column chemistry. Nobias Chelate-PA1 resin (HITACH, Japan) was used for seawater matrix removal and trace metal pre-concentration (Biller & Bruland, 2012; Sohrin et al., 2008). Recovery yields are summarized in Table 3. 5 mL eluant from each sample was analyzed for a suite of trace metals (Mn, Fe, Co, Ni, Cu, Cd, Pb) by HR-ICPMS (Thermo Element XR). Samples were introduced into the instrument with a peristaltic pump at a flow rate of ~120 uL min^-1 and passed through an ESI-PC3 Peltier cooled spray chamber before entering the torch. Sample and gas flow rates were optimized for each run; values were 0.75-0.80 ml min^-1 and 0.20-0.24 ml min^-1, respectively. Nickel sample and skimmer cones (Spectron) were used to reduce instrumental blank and memory effects. In, Y and Sc were added to each sample as internal standards for calibrating sensitivity shift of the instrument.