Journal of Geophysical Research: Biogeosciences RESEARCH ARTICLE Fe and Nutrients in Coastal Antarctic Streams: Implications 10.1029/2017JG004352 for Primary Production in the Ross Sea Key Points: Sydney Olund1,2 , W. Berry Lyons1 , Susan A. Welch1 , and Kathy A. Welch2 • Two Antarctic streams supply an average annualflux of 240 moles of 1School of Earth Sciences, Ohio State University, Columbus, OH, USA,2Byrd Polar Research Center, Ohio State University, filterable Fe to phytoplankton communities in the coastal Ross Sea Columbus, OH, USA • The average stoichiometry of these coastal streams was N3P1Si100Fe0.8, suggesting an important source of Fe AbstractThe Southern Ocean (SO) has been an area of biogeochemical interest due to the presence of and P relative to Si and N macronutrients (N, P, and Si) but lack of the expected primary production response, which is thought to be • Climate change and subsequent melt in this region is expected to increase primarily due to Fe limitation. Because primary production is associated with increased drawdown of the Feflux from these terrestrial, atmospheric CO2, it is important to quantify thefluxes of Fe and other nutrients into the SO. Here we present coastal streams data from subaerial streams thatflow into the Ross Sea, a sector of the coastal SO. Water samples were collected in the McMurdo Dry Valleys, Antarctica, and analyzed for macronutrients and Fe to determine the Supporting Information: potential impact of terrestrial water input on the biogeochemistry of coastal oceanic waters. The •Supporting Information S1 physiochemical forms of Fe were investigated through analysis of three operationally defined forms: •Table S1 •Table S2 acid-dissolvable Fe (nofiltration),filterable Fe (<0.4μm), and dissolved Fe (<0.2μm). The combined average •Table S3 flux from two McMurdo Dry Valley streams was approximately 240 moles offilterable Fe per year. The •Table S4 dissolved fraction of Fe made up 18%–27% of thefilterable Fe. The stream data yield an averagefilterable •Table S5 stoichiometry of N3P1Si100Fe0.8, which is substantially different from the planktonic composition and Correspondence to: suggests that these streams are a potential source of Fe and P, relative to N and Si, to coastal phytoplankton S. Olund, communities. While the Feflux from these streams is orders of magnitude less than estimated eolian and [email protected] iceberg sources, terrestrial streams are expected to become a more significant source of Fe to the Ross Sea in the future. Citation: Olund, S., Lyons, W. B., Welch, S. A., & Welch, K. A. (2018). Fe and nutrients in coastal Antarctic streams: Implications 1. Introduction for primary production in the Ross Sea. 1.1. Rationale Journal of Geophysical Research: Biogeosciences,123, 3507–3522. https:// It is important to understand nutrient distribution in the ocean because primary production can lead to CO2 doi.org/10.1029/2017JG004352 uptake and the subsequent sequestration of carbon via burial in marine sediments (Marinov et al., 2008; Moore et al., 2013). The distribution of nutrients in Earth’s oceans is primarily driven by ocean circulation, Received 29 DEC 2017 Accepted 11 NOV 2018 upwelling, and proximity to terrestrial sources, and these factors have created a high macronutrient (N and Accepted article online 23 NOV 2018 P) environment in the surface waters of the Southern Ocean (Moore et al., 2013). Typically, nitrate is the Published online 12 DEC 2018 limiting nutrient in the surface waters of the ocean (middle and low latitudes), and because of this, high-nitrate waters are expected to have high phytoplankton biomass (Moore et al., 2013). Work over the past decade has clearly demonstrated, however, that the micronutrient (Fe) is very important throughout various ocean locations in controlling marine primary production (Tagliabue et al., 2017). The Southern Ocean is characterized ashigh-nutrient low-chlorophyll, where primary productivity is limited by Fe instead of by macronutrients (Cassar et al., 2007; Martin et al., 1990; Moore et al., 2013). One of the earliest investigations into Fe limitation in the Southern Ocean compared the geochemistry of the productive coastal waters of Antarctica to the lower productivity offshore waters and revealed that Fe concentrations in the coastal waters were 50–60 times higher than the less productive offshore waters (Martin et al., 1990). The Fe-rich coastal waters produced 30 times more Cfixation (Martin et al., 1990). Additionally, a number of Fe fertilization experiments have shown that the uptake of CO2 increases substantially with increased availability of Fe (e.g., Watson et al., 2000). It was noted, however, that such fertilization experiments only yielded a temporary biological response to increased Fe and that transport offixed carbon to depth is required for long-term carbon sequestration from this enhanced primary productivity (Watson et al., 2000). Several studies have confirmed the earlier investigations on the importance of Fe in helping to drive primary production in Antarctic waters (Arrigo & van Dijken, 2003; ©2018. American Geophysical Union. Bertrand et al., 2015; Cassar et al., 2007; De Jong et al., 2012; Fennel et al., 2003; Smetacek et al., 2012; All Rights Reserved. Tagliabue & Arrigo, 2005), as well as having an impact on the structure of the pelagic ecosystem (Hare OLUND ET AL. 3507 Journal of Geophysical Research: Biogeosciences 10.1029/2017JG004352 Table 1 et al., 2007; Meyerink et al., 2017). More recently, a review of metal uptake Previously Calculated Fe Fluxes From Various Sources to the SO by phytoplankton throughout the global oceans showed that diatoms Feflux to SO from the Southern Ocean have a lower Fe quota than some diatoms Source (Gg/year) isolated from coastal and oligotrophic environments, which may be a Eolian deposition (Raiswell et al., 2016) ~1 strategy that has developed to cope with low Fe concentrations Eolian wet deposition (Edwards & Sedwick, 2001) 2.9 × 10 1 (Twining & Baines, 2013). Icebergs (Raiswell et al., 2016) 180–1,400 Subglacial meltwater (Death et al., 2014) 9–90 Due to this importance of Fe in influencing the biologicalfixation of CO2, MDV Terrestrial streams (Lyons et al., 2015) 1.2 × 10 2 several sources of Fe to the Southern Ocean have been investigated and Maritime Antarctic terrestrial streams dFe 1–100 theirfluxes estimated. To date, the largest estimated contributions come Maritime Antarctic terrestrial streams aFe 100–1,000 from icebergs (Raiswell, 2011). A recent study demonstrated increased (Hodson et al., 2017) chlorophyllalevels in the wake of a 50-km iceberg (termedgiant icebergs; Note. The eolian and iceberg sources from Raiswell et al. (2016) quantify Duprat et al., 2016). Other significant sources of Fe include eolian dust that which passes through a 0.2-μmfilter and is extractable with a buf- fered ascorbate solution. The eolian source from Edwards and Sedwick deposition (Edwards & Sedwick, 2001), diffusion from shelf (2001) is that which passes through a 0.2-μmfilter. The subglacial melt- sediments/porewaters (de Baar et al., 1995), convection and mixing from water source of Fe is that which passes through a 0.45-μmfilter. The deeper waters (Marsay et al., 2017; Tagliabue et al., 2014), hydrothermal MDV stream source of Fe is that which passes through a 0.45-μmfilter. MDV = McMurdo Dry Valleys; SO = Southern Ocean. sources (Tagliabue et al., 2010), the melting of sea ice (Bhattachan et al., 2015; de Jong et al., 2013; Winton et al., 2016), and subglacial melt direct from the continent (Death et al., 2014). Several studies have argued that the occurrence of offshore phytoplankton blooms are due to inputs from shelf sediments (Dulaiova et al., 2009; de Jong et al., 2013) or subglacial melt (Gerringa et al., 2012), and these sources were shown to provide the necessary Fe for the bloom to proceed. Clearly from the summation of all these studies, there are multiple sources of Fe to the Southern Ocean, and there is a need to better quantify all the contributions of Fe as well as determine the bioavailability of these sources (Tagliabue et al., 2017). Recent work has demonstrated that subaerial streamflow from ice-free areas of Antarctica, such as the McMurdo Dry Valleys (MDV), is a potential source offilterable Fe (<0.4μm, termed fFe) to the Ross Sea, with the potential to provide a mean of 0.19μmol fFe/L (Lyons et al., 2015). While this concentration, when converted to a yield on a per area basis, is small compared to other estimated sources of iron into the Southern Ocean (Table 1), climate warming, and subsequent cryospheric loss and glacier retreat could create more ice-free area (Lee et al., 2017). In addition, this increased melt could enhance streamflow, leading to a largerflux of Fe from this source in the future (Lyons et al., 2015). Recent work in other ice-free areas of the Antarctic has also clearly shown that glacier meltwater streams can contribute significantly high concentrations offilterable Fe (i.e., dissolved, colloidal, and nanoparticulate) to the marine environment surrounding the continent, enhancing the coastal water Fe concentrations (Hodson et al., 2017; Kim et al., 2015; Monien et al., 2017). This study aims to further quantify the terrestrial stream input of Fe to the McMurdo Sound. The work presented builds from the work of Lyons et al. (2015) and specifically seeks to 1. identify the physiochemical form of Fe in four MDV streams that discharge directly into the coastal Ross Sea. It is important to distinguish between the various Fe fractions—particulate, colloidal, nanoparticulate, and dissolved—due to the relationship between physiochemical form and potential bioavailability. This is complicated by the fact that the standard operational definitions of these fractions do not truly separate each chemical form and that different operational definitions occur in the literature (Raiswell, 2011; Raiswell & Canfield, 2012). In this study, we defined three fractions of Fe: a. aDFe—acid dissolvable Fe, which is unfiltered dissolvable Fe (as defined by acidification with dilute HCl for 60+ days; b.