Títulos de deuda

Seawater Fe measurements were made at four depths: 0, 5, 10 and 15 m. Although not significantly different, the mean dFe concentrations were slightly lower at the surface than at depth with concentrations starting at 2.0 ± 0.44 nM at 0 m and increasing consistently to 2.9 ± 0.54 nM at 15 m depth, approximately an order of magnitude lower than the brine-volume- normalised SI5 dFe concentration. Interestingly, these values are within the same order of magnitude as some of the surface values observed during a pack ice study tens -hundreds of kilometres from the coast (van der Merwe et al., in Press). The consistency of the observed concentrations in dFe, regardless of the proximity to Fe sources such as the continental shelf and coast around Antarctica, may indicate a dFe limit for seawater, possibly imposed by the kinetics of Fe in seawater (i.e. the rapid utilisation and conversion to the biogenic particulate fraction including Exopolysaccharide particle scavenging coupled with the short residence time of the dFe fraction in oxygen rich sea water), although further investigation is required to confirm this.

For the TDFe fraction, we observed a continual increase in concentration in the surface seawater over the first 4 stations (SD 14, 17, 20 and 23). On SD 26, we see a decrease to 14 nM of TDFe, likely resulting from reduced input from the melting sea ice above due to steady (rather than increasing) brine volume fractions in the sea ice. This is most evident in SI2 and is likely driven by the cooling of air temperature between SD 18 and 24. During the final station we observed a rapid increase to 48 nM TDFe in seawater as a result of the highest air temperatures recorded and associated brine volume fraction (appendix table). Thus, it is apparent that the air temperature is controlling the ice melt and associated TDFe flux into the underlying sea water, which in this case is primarily in the particulate form. Similarly, the mean TDFe concentration over the whole water column shows an overall increase between the first and last stations of 26 nM (appendix table).

To determine if the sea ice alone contains enough TDFe to account for the 26 nM flux into seawater, we estimated the absolute TDFe in sea ice by taking the difference between the initial and final amount of TDFe in each sea ice section. Linear interpolation with ice depth then enabled an estimate of the TDFe flux from the sea ice into the seawater during the time series which equalled 419 µmol/m2. This flux estimate averaged over the water column below (melted volume:volume) equalled 28 nM. Thus, the TDFe flux from the overlying sea ice can entirely explain the observed enrichment in the underlying 16 m of sea water. The same calculation for the dFe fraction resulted in a flux estimate of 25 µmol/m2 released into the water column during the study period. This dFe flux averaged over the 16 m water column results in an enrichment of 1.6 nmol/L and thus does not completely explain the observed enrichment.

The TDFe flux estimate equates to 419 µmol /m2 of similar coastal fast ice during the study period (i.e. 28 days). Thus, the TDFe flux observed was an order of magnitude higher than

during a 31 day study period in pack ice in the Western Weddel Sea (Lannuzel et al., 2008). Conversely, the dFe fraction was only twice that observed during Lannuzel et al (2008). Although Lannuzel et al (2008) sampled later in the season (November 29 – December 30) and therefore may have missed some of the initial Fe stock, our result indicates the potential of coastal fast ice as a significant TDFe reservoir. The similar dFe flux and vastly different TDFe flux observed between these two studies indicates, either, some controlling factor on the concentration of dFe and not TDFe, and/or that the sources of these two fractions are independent which has been suggested by prior research (Grotti et al., 2005; Lannuzel et al., 2008; Lannuzel et al., 2010; van der Merwe et al., in Press). A likely explanation is that the source is primarily in the particulate fraction (from sediment derived deep water) and this is only converted to the dissolved fraction upon in situ biogenic and photooxidative transformations and remineralisation.

Although there is growing evidence that sea ice entrained Fe may be one of the most bioavailable forms due to abundant organic complexation coupled with photo-oxidation either in situ or upon release into strongly stratified melt waters (Kim et al., 2010; Rijkenberg et al., 2008; Steigenberger et al., 2009; Tagliabue et al., 2009b), much of this Fe will be supplied in the particulate form when released from melting sea ice, and therefore, not readily available for phytoplankton uptake. Thus, the majority of this Fe will not be delivered to the Fe limited waters further off shore without significant heterotrophic remineralisation and photooxidation to keep it in solution. However, it is possible that regional circulation patterns within the water column could allow the northward movement of this Fe enriched fast ice and associated sea water, past the Antarctic Divergence via diversions in the East Wind Drift that carry ice northward (Worby et al., 1998). These divergences have been observed within the vicinity of the study site at 95 and 130˚ E and may supply large quantities of Fe enriched seawater northwards.

Thus, and with the above caveat in mind, if this amount of Fe was delivered into typical Fe limited Southern Ocean surface waters (which often have ~0.1 nmol L-1 ambient dFe concentration in summer, (Bowie et al., 2001)) it is sufficient to raise 419 m3 of sea water by 1 nmol L-1 per m2 of similar coastal sea ice thereby typically alleviating limitation.

5.5.

Conclusion

We conducted a series of physical and biogeochemical observations at a coastal fast ice station in a period of rapid warming and associated ice melt during late austral spring. The air temperature was clearly associated with changes in brine volume fraction. This was particularly evident within the upper interior ice, which was insulated from rapid changes from above and below. Macronutrient profiles revealed the nitrate deficient state of the interior ice; this is most obvious in the brine concentrations which show a large excess of silicate and a deficit of nitrate. In the basal ice, relative to sea water, nitrate and silicate were drawn down through time, but due to resupply from below, most likely did not lead to a limiting condition. POC and Chl a showed a degree of variability through the time series, reflecting inter-core heterogeneity, difficulties of complete sample homogenisation and rapid melting and ice mass loss in the skeletal layer at the ice water interface. Dissolved Fe is highly diffuse and readily transferred from the surface/interior to the basal sea ice layers as the ice melting progressed. In contrast, pFe did not show this clear decreasing trend and correlated with POC and Chl a distributions. Indeed, the factor analysis revealed that 53% of the total variance of all the variables could be accounted for by one factor most closely correlated with POC. Thus, POC and all fractions of Fe measured were closely correlated. Furthermore, the macronutrients (particularly phosphate and nitrate) were distinct (factor 2, 27.6 %), most likely due to active drawdown by the sea ice algal assemblage. The non-refractory plFe fraction did decrease through time in most sea ice sections and was deposited into the underlying seawater. Using TDFe concentration, we calculated the

decrease in sea ice entrained Fe through time as the melting progressed. This deficit in Fe was released into the underlying water column and can completely explain the seawater enrichment observed. Over the 28 days of sampling, two distinct mean-air-temperature warming events were observed of -12.1 to -1.3 ˚C and then -6.4 to 0.8 ˚C. This resulted in the release of 419 µmol TDFe and 29 µmol dFe per m2 of sea ice from this coastal fast ice station into the underlying water column. This TDFe enrichment is an order of magnitude higher and the dFe is double prior observations in pack ice in the Western Weddell Sea and indicates the potential of coastal fast ice as a significant Fe reservoir. Given the trace quantities of Fe required by typical Antarctic diatoms to bloom, this represents a fertilisation potential for 419 m3 of Fe limited surface Southern Ocean seawater with TDFe and 29 m3 with dFe, per m2 of similar coastal fast ice.

5.6.

Acknowledgments

We gratefully acknowledge the management, laboratory manager and support staff at Casey Station (Australia). We would like to thank Andy Cianchi, Simon Cross and Mick Stapleton for field support to gain access to the sampling site. This work was partially funded by the Australian Government’s Cooperative Research Centres Program through the Antarctic Climate and Ecosystems Cooperative Research Centre (ACE CRC). This project was sponsored by an Australian Antarctic Science grant (#3026). We would like to thank Dr Ashley Townsend and the Central Science Laboratory (CSL) at the University of Tasmania for invaluable mentoring and support with the ICP-MS analyses. Total nitrogen, carbon and hydrogen was determined by Dr Thomas Rodemann at the Central Science Laboratory, University of Tasmania. The analysis for macronutrients was made by the CSIRO Hydrochemistry team, Hobart.