1. Elevated suspended particulate Hg species concentrations, lower K d in the upwelling compared to the open ocean
Profiles of THg for SSF and LSF are presented in Fig. 2. Particulate THg concentrations are generally highest with the mixed layer of 50m and slightly attenuate with depth. The THg concentration is much higher in SSF (28 to 114 fM) and has a wider concentration range than LSF (2 to 25 fM) across all stations due to higher concentrations of SSF in the water column (Table 1). The upwelling regions exhibited higher THg than the open ocean. In the upper ocean (above 200 m), the ratio of particulate THg to dissolved THg (presented in Adams et al. (2024, submitted), DOI) averaged at 14.55% (range: 3–47%), which is higher than the ratios reported in the Equatorial Pacific16, indicating particles are an important phase of Hg in the coastal upper ocean. Particulate MMHg data is mostly available in the upwelling regions (Table 1, Figure S1), with range of 1.25–19.32 fM in the SSF and 0.11–2.12 fM in the LSF. This information is relevant to subsequent analysis on Kd and MMHg ratios.
The calculated Kd of THg (C-based) in the SSF was not consistent across the upwelling region and open ocean, with statistically higher values (t-test; p-value = 0.016 < 0.05) found in the oligotrophic region (log Kd, upwelling: 6.59 ± 0.52; oligotrophic: 7.21± 0.26). This is in line with our previous finding that the open ocean generally exhibits higher Kd values when compared to the coastal regions due to the so-called particle concentration effect.30 Additionally, Kd in the SSF is similar to that in the LSF (LSF log Kd, upwelling: 6.61 ± 0.38) (t-test; P-value = 0.42 > 0.05). This similarity indicates that Kd is a consistent value present in the SSF and LSF. It also indicates aggregation and disaggregation are frequently mixing the SSF and LSF pools at fast rates relative to the residence time of each pool above 200 m depth. Compared to THg, the Kd for MMHg in the upwelling region is approximately one order of magnitude lower, with a log Kd of 5.68, indicating that MMHg exhibits lower particle activity compared to inorganic mercury.
2. CCE upwelling increases Hg sinking flux and evasion
Hg sinking fluxes were examined from sinking particles, exhibiting a spatial pattern of decreasing from the coastal upwelling region to the open ocean. Unlike C mass fluxes, THg sinking fluxes did not attenuate with depth in each cycle. They had a maximum at 150 m with a value highest at Cycle 1 (2879 pmole m− 2 d− 1) and lower at Cycle 2 and 3 (1369, 1478 pmole m− 2 d− 1, respectively; Table 2). We considered the possibility of lateral injection of particles from the margin as a contributing factor, but beam attenuation analysis (Figure S2) did not reveal any mid-water anomalies in particles, suggesting that this is not a likely explanation. The high sinking fluxes of Hg species observed in the CCE, which are one order of magnitude higher than those reported in the Equatorial Pacific,16 may result from high productivity in the CCE. Indeed, we observed higher C export in the CCE compared to the Equatorial Pacific upwelling region (e.g., 120 mg m− 2 d− 1).
Inspired by the unexpected high Hg sinking flux, we developed a box model to study Hg dynamics incorporated upwelling in the CCE. Modeled Hg budgets in the upper ocean within the CCE upwelling region are presented in Fig. 3a and Table S2. In general, all variables reach a steady state in about 40 days (Figure S3), indicating a relatively short residence time of Hg in the mixed layer. The high similarity between simulated and field observed values indicates the robustness of the model (Table 3). Hg sinking flux is the largest removal for Hg (95%). In contrast, upwelling supplies 42% dissolved and particulate Hg species to the mixed layer. As atmospheric deposition and evasion accounts for a small part of the gain (14%) and loss (5%) of Hg, the estimated removal fluxes from the mixed layer exceed inputs of Hg, suggesting that Hg is imported from the neighboring ocean by horizontal advection.
To better understand the effect of upwelling on Hg sinking flux, we compared the model outputs under two scenarios with distinct upwelling velocity (1.12 vs. 0 m d− 1). When removing the upwelling input, the model predicts a decreased sinking flux by 40% (Fig. 5b), indicating that upwelling significantly enhances sinking flux. The field data we used to parameterize NPP in our model suggested a baseline NPP of 274 mgC m− 2 d− 1 in the absence of upwelling. This baseline, however, has been observed to be lower in previous CCE investigations.31 Our estimation of the impact of upwelling on sinking flux is therefore a conservative one, yet still significant.
Usually, increased upwelling introduces additional nutrients into surface water, resulting in enhanced primary production. In the CCE, primary production was reported to have a significant positive relationship with export.31 We do observe higher NPP and export in the upwelling compared to the open ocean during our cruise (Table S3). Thus, we would anticipate the high Hg export in the CCE upwelling region due to the high affinity of Hg to POC.30 Gravity sinking is not the only way to remove Hg from the surface, in addition, high subduction rates of seawater at the front may also contribute to total Hg export, which cannot be detected by sediment traps.32 This suggests that the total Hg removal from surface may be underestimated. However, subduction is typically more significant in offshore, oligotrophic regions than in upwelling zones.32 Therefore, our estimated Hg sinking flux remains valid and has strong positive relationship with upwelling.Similar observations of high Hg sedimentation related to sorption and scavenging was found in the other regions with upwellings (e.g., Antarctic, Peruvian upwelling)17–19.
Furthermore, the comparison of results from the two scenarios reveals that upwelling not only influences Hg sinking flux but also plays a significant role in other Hg processes such as Hg evasion and inventory. Notably, evasion exhibits a remarkable response to varying upwelling velocity. Our simulations indicate a marked 45% increase in Hg evasion under upwelling condition, consistent with previous observations within CCE that augmented evasion fluxes facilitated by upwelling-induced cyclonic eddies.5 Because biotic reduction rate is prevalent in highly productive regions, contributing to elevated Hg evasion rates.33 While it is expected to observe a higher THg inventory brought by upwelling, we modeled a contrary decrease of dissolved THg associated with upwelling. With rapid response of particle scavenging and Hg evasion, the dissolved Hg is swiftly absorbed by the particles and undergoes reduction to Hg0 followed by subsequent re-emission.
The Hg cycle within the CCE upwelling region exhibits varying degrees of sensitivity to the parameters (Fig. 7). Kd profoundly impacts on various Hg processes including sinking flux, evasion, and dissolved Hg2+ inventory. For instance, a 7% increase of in log Kd results in a 131% increase in sinking flux, a 32% decrease in evasion, and a 24% decrease in dissolved Hg pool. Conversely, wind speed demonstrates a relatively minor effect on the overall Hg cycle, primarily influencing evasion. Specifically, a 6-fold increase in wind speed leads to an 18-fold increase in evasion, while the dissolved Hg pool and sinking flux exhibit a subtle response with a decrease by 4%. It is worth noting that while the various parameters have a crucial role in determining Hg cycles, they do not alter the fundamental premise that upwelling can enhance sinking flux.
3. SSF is a favorable site for MMHg compared to the LSF
As noted above, particles can provide anaerobic environment for the Hg methylation, however, little is known is about the preference of MMHg for particles of different size spectrum. While it has been reported that LSF can sustain anaerobic environment for sulfate-reducing within wider depth.8 Our findings indicate a significantly higher average MMHg to THg ratio in the SSF at 19.17% compared to the LSF at 11.69%, as determined by a t-test with a p-value = 0.016. This suggests that the SSF may serve as a more favorable micro-environment for MMHg production or retention. Additionally, our analysis of C data (Table 4, Figure S4) reveals higher (not significantly different) MMHg to C ratios in the SSF compared to the LSF, providing further evidence of the SSF's conducive nature for MMHg.
Interestingly, MMHg/THg ratios averaged at 2.15% in sediment traps were much lower compared with SSF and LSF. This discrepancy in ratio could be attributed to the biological aggregation process.34, 35 It is known that particles are often consumed by zooplankton,36 which can assimilate MMHg from their diet. As zooplankton retain MMHg in their bodies, the sinking particles they produce (fecal pellets) should be expected to contain less MMHg relative to THg as we observed.
4. Implication of intensified coastal upwelling
Coastal upwelling is predicted to intensify as a consequence of global warming14, 37. Although, our model suggests that coastal upwelling results in an increase in evasion and decrease in dissolved THg pool in the upper ocean, consequently reducing the potentially bioavailable Hg2+ that may be converted to MMHg, however, with a significant downward export of Hg would probably result in more in-situ microbial production of MMHg in deeper water under lower oxygen conditions. Increasing methylated Hg generated in the deep water is then transported to the mixed layer by upwelling, which leads to a higher particulate MMHg/THg ratio in the mixed layer. Such dynamics could potentially promote Hg level in the local food chain through bioaccumulation. Further field investigations are needed to fully comprehend the intricate dynamics of methylmercury (MMHg) under the influence of upwelling.