Upwelling Enhances Mercury Particle Scavenging in the California Current System

Coastal upwelling supplies nutrients supporting primary production while also adding the toxic trace metal mercury (Hg) to the mixed layer of the ocean. This could be a concern for human and environmental health if it results in the enhanced bioaccumulation of monomethylmercury (MMHg). Here, we explore how upwelling influences Hg cycling in the California Current System (CCS) biome through particle scavenging and sea-air exchange. We collected suspended and sinking particle samples from a coastal upwelled water parcel and an offshore non-upwelled water parcel and observed higher total particulate Hg and sinking flux in the upwelling region compared to open ocean. To further investigate the full dynamics of Hg cycling, we modeled Hg inventories and fluxes in the upper ocean under upwelling and non-upwelling scenarios. The model simulations confirmed and quantified that upwelling enhances sinking fluxes of Hg by 41% through elevated primary production. Such an enhanced sinking flux of Hg is biogeochemically important to understand in upwelling regions, as it increases the delivery of Hg to the deep ocean where net conversion to MMHg may take place.


Model Variable
Air-sea exchange of elemental Hg (F eva ; ng L -1 h -1 ) is modeled as in in Soerensen et al. 1 ; F eva is set equal to: Where [Hg 0 ] sea is the Hg 0 concentration in surface seawater (pM), [Hg 0 ] atm is the Hg 0 concentration in the air (also in pmole/L units), H is the unitless Henry's Law constant and a function of temperature and salinity (Andersson et al., 2008)  2 and k eva is the mass transfer coefficient (also called a "piston velocity"), calculated using the Nightingale et al. formulation. 3mospheric deposition of Hg 2+ (F dep ; ng L -1 h -1 ) is estimated using the GEOS-Chem The sinking flux of Hg (F sink : ng L -1 h -1 ) is modeled by multiplying particulate Hg by a particle deposition rate.The particle deposition rate (v dep : m 2 d -1 ) is calculated as export (mgC m - 2 d -1 ) divide by particulate organic carbon (POC; mgC m -3 ) (Table S1).
The upwelling flux of Hg (F upwelling ; ng L -1 h -1 ) is estimated by multiplying the vertical upwelling velocity of water by THg concentration in the intermediate seawater just below the model depth boundary.Upwelling velocity is estimated by dividing upwelling index (UI; m 2 d -1 ) by upwelling cross-shore length (UL; m) (Table S1).
Hg export to and import from the neighboring ocean (F ex , unit: ng The total Hg 2+ and Hg 0 concentration (ng L -1 ) in the seawater were calculated by integrating the rate of change in fluxes over time (h).Particulate Hg 2+ (ng L -1 ) was calculated using an equation based on the partitioning coefficient (K d ) equation 5,6 .k red-Hg(II) and k ox-Hg0 are the pseudo-first order reduction rate constant of Hg(II) and oxidation rate (h -1 ). 4

Estimate of Atmospheric Dust Deposition
We estimated the Hg flux from atmospheric dust to the ocean, using Hg/Al ratio (0.5 * 10 -6 g/g) in the south Pacific 7 , dust flux (18.4 mg/m 2 /d) and Al concentration (14.1 * 10 2 mol /g) in dust in the southern CCS 8 .The estimated Hg flux from dust is 350 pg/m 2 /d.The particulate Hg in dust is ~19 ng/g.Although we do not have total particle mass, if we assume that total carbon mass approximately accounts for 20%, then the rough Hg concentration in ng/g, calculated from Hg/C in our study, is 290 ng/g.Therefore, the Hg from dust would only account for 6.5% of total particulate Hg.This suggests that dust is not a big source to particulate Hg in our study region.

Model Outputs
Table S3 Simulated Hg budgets showing median values for upwelling and non-upwelling scenarios.A '0' in non-parametric test indicates no significant difference between two scenarios, while a "1" indicates significant difference. /doi.org/10.5281/zenodo.5075847).
is simplified and expressed as water mass exchange rate multiplied by the THg concentrations in the "upstream" water mass (flux in), or the THg concentration within the model domain (flux out).We follow the method of Liu et al. to simplify the water mass exchange rate by dividing domain water volume (V) by water residence time (rt, unit: d).4

Figure S1
Figure S1 Profiles of carbon (C) mass (mol L -1 ), THg to C ratio, and MMHg to C ratio (pmol mol −1 ) distribution with depth (m) in upwelling and open ocean.Panel (a), (b), and (c) are for the C mass, THg/C ratio, and MMHg/C ratio in SSF, respectively; panel (d), (e), and (f) are for the C mass, THg/C ratio, and MMHg/C in LSF.Error bars represent standard deviations.

Figure S2
Figure S2Monte Carlo simulation plots for selected variables.