Use of Fe isotopes to examine Fe cycling in Boreal Shield lakes: implications for cyanobacterial bloom development and ancient biogeochemical cycling

Abstract

Stable iron (Fe) isotopes have only been used recently to explore biogeochemical Fe processes in terrestrial and aquatic ecosystems. The δ⁵⁶Fe values in natural samples range from −4‰ to 4‰. An increasing number of studies have focused on Fe isotopes in marine settings. However, few studies have focused on lake systems and only in extreme environments such as meromictic ferruginous lakes. This thesis has applications in two distinct fields, freshwater eutrophic research and early Earth evolution. Phytoplankton blooms have been studied extensively in lake ecosystems. Despite the fact that macronutrients phosphorus and nitrogen are commonly known as dominant factors controlling the blooms, the critical role of micronutrient Fe has been proposed in recent studies. However, current understanding of the sources and availability of Fe for phytoplankton blooms and Fe cycling in lakes in general remains limited. Although the mechanisms that result in the deposition of banded iron formation in the Archean ocean remain a subject of debate, it is proposed that photoferrotrophy played a significant role in the oxidation of Fe(II) in anoxic Archean oceans. Studies investigating the occurrence and significance of photoferrotrophy have primarily focused on meromictic and ferruginous lakes. However, such lakes are thought to be rare on a global scale. To explore the potential application of stable Fe isotopes for tracing biogeochemical Fe cycling in Boreal Shield lakes, δ⁵⁶Fe values of dissolved, particulate, and sediment Fe were measured in two small dimictic Boreal Shield headwater lakes: manipulated eutrophic Lake 227, with annual cyanobacterial blooms, and unmanipulated oligotrophic Lake 442. Within these small lakes, the range in δ⁵⁶Fe is large (ca. −0.9 to +1.8‰), spanning more than half the entire range of natural Earth surface samples. Two layers in the water column with distinctive δ⁵⁶Fe of dissolved (DFe) and particulate Fe (PFe) were observed in both lakes, despite large differences in trophic states. During cyanobacterial blooms in Lake 227, selective uptake of isotopically light Fe modifies δ⁵⁶Fe, resulting in Δ⁵⁶Fe dis-part of up to 1‰ between dissolved and particulate Fe in the epilimnion while little fractionation was observed in the epilimnion of Lake 442. In the anoxic layers in both lakes, upward flux from sediments dominates the dissolved Fe pool with an apparent Δ⁵⁶Fe dis-part of −2.2 to −0.6‰. Large Δ⁵⁶Fe dis-part and previously published metagenome sequence data suggest active Fe cycling processes in anoxic layers, such as microaerophilic Fe(II) oxidation or photoferrotrophy, could regulate biogeochemical cycling. Large fractionation of stable Fe isotopes in these lakes provides a potential tool to probe Fe cycling and the acquisition of Fe by cyanobacteria, with relevance for understanding biogeochemical cycling of Earth’s early ferruginous oceans. To further explore the effect of phytoplankton blooms on biogeochemical Fe cycling in the epilimnion and metalimnion of L227, seasonal variations in Fe concentrations, Fe speciation and Fe isotope compositions of dissolved Fe (DFe), particulate Fe (PFe) and total Fe (TFe) were reported in the oxic layers of L227 and L442. Biological uptake of Fe by phytoplankton results in a decrease in DFe, an increase in PFe and a decrease in TFe in the oxic epilimnion and metalimnion during the blooms. The prevalence of Fe(II) in the particulate Fe coupled with the depletion of Fe(II) in the dissolved Fe during the blooms suggest that diazotrophic cyanobacteria and chlorophytes in L227 likely utilize dissolved inorganic Fe(II) rather than organic complexed Fe(III), highlighting the importance of Fe(II) during iron acquisition. Uptake of isotopically light Fe by phytoplankton results in a diagnostic change in the dissolved and particulate δ⁵⁶Fe in the oxic layers. As a result, dissolved δ⁵⁶Fe is more positive than particulate δ⁵⁶Fe in the oxic layers of L227, with Δ⁵⁶Fe dis-part of 1.02 ± 0.33‰ (2σ) during the peak of the first bloom and Δ⁵⁶Fe dis-part of 0.56 ± 0.05‰ (2σ) during the second bloom. The difference in Δ⁵⁶Fe dis-part observed during the first bloom and the second bloom suggests that uptake of Fe by chlorophytes might produce distinct Fe isotope fractionation compared to diazotrophic cyanobacteria, possibly due to the differences in their uptake kinetics, the specific Fe-uptake mechanisms and associated enzymatic processes. The consistency in TFe concentration and Fe isotope composition of TFe in the epilimnion throughout the first and second blooms suggest that the settling flux of particulate Fe is limited during the blooms, likely due to efficient retention by phytoplankton. In contrast, during the decline of the bloom, the settling of isotopically light particulate Fe(II) associated with biomass to profundal sediment caused total δ⁵⁶Fe in the epilimnion to become isotopically heavier. During the peak of the first bloom, the large Δ⁵⁶Fe dis-part, the more negative particulate δ⁵⁶Fe, and the more particulate Fe concentrations in the metalimnion compared to what has been observed in the epilimnion suggest that cyanobacteria could migrate to the upper zone of the anoxic hypolimnion to uptake Fe under Fe-limited conditions. Modelling results showed that isotopically light PFe and TFe measured in the metalimnion during the peak of the first bloom are best explained by the migration of cyanobacteria below the redox boundary for the uptake of dissolved Fe(II). The observed seasonal variation in Fe concentration, speciation and isotope compositions associated with the progression of two annual blooms show great potential in understanding the role of Fe in the formation of cyanobacterial bloom and determining the availability and source of Fe for cyanobacterial uptake. To explore the Fe processes and Fe sources in the anoxic hypolimnia of Boreal Shield lakes, seasonal variation in Fe concentrations and Fe isotope compositions of dissolved Fe and particulate Fe in the anoxic hypolimnia along with the Fe concentrations and Fe isotope compositions of porewater Fe and bulk sediment Fe in bottom sediment cores were reported in four Boreal Shield lakes: one eutrophic Lake 227 and three oligotrophic Lake 221, Lake 304 and Lake 442. In the oxic layers of all lakes, dissolved δ⁵⁶Fe is more positive than particulate δ⁵⁶Fe, likely modified by a combination of multiple processes including biological uptake, photochemical reduction of particulate Fe(III), complexation of dissolved Fe and dissolved organic matter, photolysis and microbial decomposition. The observed Δ⁵⁶Fe dis-part in the oxic layers of all lakes ranges from 0.06 ± 0.09‰ (2σ) to 1.02 ± 0.33‰(2σ), with the higher values likely arising from biological uptake. In contrast, the δ⁵⁶Fe values of dissolved and particulate Fe are reversed in the anoxic hypolimnia of all lakes, with negative Δ⁵⁶Fe dis-part fractionations of −0.91 ± 0.58‰ (2σ) in L221, −0.87 ± 0.81‰ (2σ) in L227, −1.06 ± 0.26‰ (2σ) in L304 and −0.72 ± 0.25‰ (2σ) in L442. Particulate Fe from the epilimnia is likely reduced completely by dissimilatory iron-reducing bacteria at the redox boundary, supported by the similarity in δ⁵⁶Fe between dissolved Fe in the anoxic hypolimnia and particulate Fe in the epilimnia of all lakes. The highly positive δ⁵⁶Fe values of particulate Fe in the anoxic hypolimnion and the large Δ⁵⁶Fe dis-part fractionations indicate in situ active Fe cycling in the anoxic hypolimnion. The most plausible explanation for the large Δ⁵⁶Fe dis-part fractionations in the anoxic hypolimnion is microbial oxidation, possibly photoferrotrophy, supported by the 16S rRNA gene sequencing and genome-resolved metagenome sequencing analysis. An increase in the magnitude of Δ⁵⁶Fe dis-part with depth and over time was observed in L221, L227 and L304, likely due to the variation of the oxidation rates, supported by increased particulate Fe concentrations with depth and time. The high concentration and negative δ⁵⁶Fe values of porewater Fe(II) suggest that DIR dominates organic mineralization pathways in sediment cores from the bottoms of L227, L304 and L442. The DIR-produced porewater δ⁵⁶Fe is further modified by the diagenetic formation of siderite in the upper section of sediment cores and the diagenetic formation of pyrite in the deeper section. Temporal and spatial variation in Fe isotope fractionation in the anoxic hypolimnia, along with the Fe isotope signatures in the sediment, suggest active Fe cycling in four lakes, highlighting the potential of seasonally anoxic Boreal Shield lakes to serve as analogues of the late Archean ocean.