Groundwater infiltration as a potential trigger for carbonate precipitation
Oxygen isotope variability may be indicative of isotopic heterogeneity in the fluid from which these carbonates precipitated, potentially caused by mixing of multiple water sources with significantly different oxygen isotope compositions. Stream waters in the Fryxell basin flow into Lake Fryxell during the summer and have slightly heavier (0.8-1.2‰) δ18O values than the water column (Gooseff et al. 2006, Harris et al. 2007). This relative enrichment is not of a sufficient magnitude to fully explain the oxygen isotope offset observed in Lake Fryxell carbonates, nor are other documented water sources in the basin. Lake and glacier ices, surface waters, and regional evaporatively concentrated Ca-chloride brines derived from meteoric waters cannot account for the heaviest carbonate δ18O values (Figure S4) (Fountain 2014, Matsubaya et al. 1979). Carbonate δ18O values are consistent with predicted precipitation from marine fluids at ambient temperatures (Figure S4), but despite the proximity of McMurdo Sound ~5 km down valley from Lake Fryxell, there is no evidence for surface inflow to the lake.
More diverse waters may interact with lake sediments through groundwater infiltration. Lake Fryxell is connected via a talik (an unfrozen region of ground bounded by permafrost) to a regional subsurface calcium chloride-rich brine aquifer (Foley et al. 2019, Mikucki et al. 2015, Toner and Sletten 2013). The aquifer is most likely a remnant of GLW which infiltrated the ground after extensive freezing and/or evaporation created a concentrated brine (Mikucki et al. 2015, Myers et al. 2021). Although no isotopic data exist for this brine at present, the brine would likely have higher δ18O values than modern meteoric-sourced waters due to incorporation of marine fluids; seawater has previously been suggested as an explanation for carbonate18O enrichment in sediments deposited from GLW (Lawrence and Hendy 1989). Any evaporative concentration with brine formation would further enrich these values as observed in other surface waters of the McMurdo Dry Valleys (Horita 2009). The isotopic data presented here indicate that this brine could have infiltrated through the lake sediments and mixed with mat pore waters, disrupting the geochemical equilibrium and inducing short-lived precipitation of isotopically heterogeneous carbonates documented in this study.
The lack of covariation between δ18O and δ13C is also consistent with precipitation from a low-DIC calcium chloride brine, such as those which have been observed in Taylor and Wright Valleys (Dickson et al. 2013, Toner and Sletten 2013). In this case, the induction of carbonate precipitation could be the product of locally increasing calcium concentrations in the pore waters of benthic mats, counteracting the otherwise low calcium levels within the pycnocline (Lawrence et al. 1985).
Geochemical constraints on Lake Fryxell carbonates suggest that an episode of enhanced connection between existing lake water and evaporatively-modified marine or marine-influenced ground water occurred in recent decades following lake level rise and shallowing of the lake oxycline. Future work on oxygen isotope composition of Antarctic carbonates and their source waters promises new insights into the hydrology and climate history of this region, as well as the processes which induce carbonate precipitation in Antarctic lakes. Additionally, the high variability of δ18O across μm to mm in Lake Fryxell carbonates highlights the importance of high-resolution sampling in environmental reconstructions in these polar deposits; sample homogenization during bulk analysis may obscure variability important to interpret processes surrounding carbonate precipitation. Applying similar combined geochemical and petrographic characterization to Antarctic paleolake deposits is necessary to reconstruct the sources of carbonates and their significance for interpretation of terrestrial Antarctic environments.