To elucidate the role of (bio)geochemical processes that fueled iron and carbon cycling in early Earth oceans, modern environments with similar geochemical conditions are needed. As the range of chemical, physical, and biological attributes of the Precambrian oceans must have varied in time and space, lakes of different compositions are useful to ask and answer different questions. Tropical Lake Matano (Indonesia), the largest known ferruginous lake, and Lake Pavin (France), a meromictic crater lake, are the two best studied Precambrian ocean analogs. Here we present seasonal geochemical data from two glacially formed temperate ferruginous lakes: Brownie Lake (MN) and Canyon Lake (MI) in the Upper Midwest, USA. The results of seasonal monitoring over multiple years indicate that (1) each lake is meromictic with a dense, anoxic monimolimnion, which is separated from the less dense, oxic mixolimnion by a sharp chemocline; (2) below this chemocline are ferruginous waters, with maximum dissolved iron concentrations >1 mM; (3) meromixis in Brownie Lake is largely anthropogenic, whereas in Canyon Lake it is natural; (4) the shallow chemocline of Brownie Lake and high phosphorus reservoir make it an ideal analog to study anoxygenic photosynthesis, elemental ratios, and mineralogy; and (5) a deep penetrating suboxic zone in Canyon Lake may support future studies of suboxic microbial activity or mineral transformation.Plain Language Summary Earth's atmosphere acquired oxygen around 2.46-2.33 billion years ago from oxygen-producing bacteria. Researchers are interested in how elements, for instance, iron and carbon, cycled between dissolved and solid phases in the ocean before oxygen was produced. In past oceans with little oxygen, all life would have been microbial. In order to study elemental cycling on early Earth, modern water bodies where oxygen is absent in deep water are needed. Water samples were analyzed from different depths within Brownie Lake in Minneapolis, MN, and Canyon Lake in the Upper Peninsula of MI. We found that these lakes have two distinct chemical layers like past oceans: a top section that has oxygen and a bottom section that has no oxygen. The bottom sections never mix with the water on top and so remain without oxygen year-round. Brownie and Canyon Lakes have different, but applicable water chemistries as compared to well-studied field sites in Europe, Africa, and Asia. With the addition of Brownie Lake and Canyon Lake as early Earth analogs, researchers will have the potential to answer many questions about early Earth's oceans.

No abstract

Meromictic lakes with anoxic bottom waters often have active methane cycles whereby methane is generally produced biogenically under anoxic conditions and oxidized in oxic surface waters prior to reaching the atmosphere. Lakes that contain dissolved ferrous iron in their deep waters (i.e., ferruginous) are rare, but valuable, as geochemical analogues of the conditions that dominated the Earth's oceans during the Precambrian when interactions between the iron and methane cycles could have shaped the greenhouse regulation of the planet's climate. Here, we explored controls on the methane fluxes from Brownie Lake and Canyon Lake, two ferruginous meromictic lakes that contain similar concentrations (max. >1 mM) of dissolved methane in their bottom waters. The order Methanobacteriales was the dominant methanogen detected in both lakes. At Brownie Lake, methanogen abundance, an increase in methane concentration with respect to depths closer to the sediment, and isotopic data suggest methanogenesis is an active process in the anoxic water column. At Canyon Lake, methanogenesis occurred primarily in the sediment. The most abundant aerobic methane‐oxidizing bacteria present in both water columns were associated with the Gammaproteobacteria, with little evidence of anaerobic methane oxidizing organisms being present or active. Direct measurements at the surface revealed a methane flux from Brownie Lake that was two orders of magnitude greater than the flux from Canyon Lake. Comparison of measured versus calculated turbulent diffusive fluxes indicates that most of the methane flux at Brownie Lake was non‐diffusive. Although the turbulent diffusive methane flux at Canyon Lake was attenuated by methane oxidizing bacteria, dissolved methane was detected in the epilimnion, suggestive of lateral transport of methane from littoral sediments. These results highlight the importance of direct measurements in estimating the total methane flux from water columns, and that non‐diffusive transport of methane may be important to consider from other ferruginous systems.

Most manganese (Mn) enrichments in the sedimentary rock record are hosted in carbonate minerals, which are assumed to have formed by diagenetic reduction of precursor Mn-oxides, and are considered diagnostic of strongly oxidizing conditions. Here we explore an alternative model where Mn-carbonates form in redox-stratified water columns linked to calcium carbonate dissolution. In ferruginous Brownie Lake in Minnesota, USA, we document Mn-carbonates as an HCl-extractable phase present in sediment traps and in reducing portions of the water column. Mn-carbonate become supersaturated in the Brownie Lake chemocline where dissolved oxygen concentrations fall below 5 μM, and Mn-oxide reduction increases the dissolved Mn concentration. Supersaturation is enhanced when calcite originating from surface waters dissolves in more acidic waters at the chemocline. In the same zone, sulfate reduction and microaerobic methane oxidation add dissolved inorganic carbon (DIC) with negative δ 13 C. These observations demonstrate that sedimentary Mn enrichments may 1) develop from primary carbonate phases, and 2) can occur in environments with dissolved oxygen concentrations <5 μM. Primary Mn-carbonates are likely to originate in environments with high concentrations of dissolved Mn (>200 μM), and where Mn and Fe are partitioned by S cycling, photoferrotrophy, or microaerophilic Fe-oxidation. A shallow lysocline enhances Mn-carbonate production by providing additional DIC and nucleation sites for crystal growth. This carbonate model for Mn-enrichments is expected to be viable in both euxinic and ferruginous environments, and provides a more nuanced view