Title: Identification and environmental interpretation of diagenetic and biogenic greigite in sediments: A lesson from the Messinian Black Sea
Abstract: Geochemistry, Geophysics, GeosystemsVolume 15, Issue 9 p. 3612-3627 Research ArticleFree Access Identification and environmental interpretation of diagenetic and biogenic greigite in sediments: A lesson from the Messinian Black Sea Liao Chang, Corresponding Author Liao Chang Paleomagnetic Laboratory “Fort Hoofddijk,” Department of Earth Sciences, Utrecht University, Utrecht, Netherlands Research School of Earth Sciences, Australian National University, Canberra, Australian Capital Territory, AustraliaCorrespondence to: L. Chang, [email protected]Search for more papers by this authorIuliana Vasiliev, Iuliana Vasiliev Paleomagnetic Laboratory “Fort Hoofddijk,” Department of Earth Sciences, Utrecht University, Utrecht, NetherlandsSearch for more papers by this authorChristiaan van Baak, Christiaan van Baak Paleomagnetic Laboratory “Fort Hoofddijk,” Department of Earth Sciences, Utrecht University, Utrecht, NetherlandsSearch for more papers by this authorWout Krijgsman, Wout Krijgsman Paleomagnetic Laboratory “Fort Hoofddijk,” Department of Earth Sciences, Utrecht University, Utrecht, NetherlandsSearch for more papers by this authorMark J. Dekkers, Mark J. Dekkers Paleomagnetic Laboratory “Fort Hoofddijk,” Department of Earth Sciences, Utrecht University, Utrecht, NetherlandsSearch for more papers by this authorAndrew P. Roberts, Andrew P. Roberts Research School of Earth Sciences, Australian National University, Canberra, Australian Capital Territory, AustraliaSearch for more papers by this authorJohn D. Fitz Gerald, John D. Fitz Gerald Research School of Earth Sciences, Australian National University, Canberra, Australian Capital Territory, AustraliaSearch for more papers by this authorAnnelies van Hoesel, Annelies van Hoesel Department of Earth Sciences, Utrecht University, Utrecht, NetherlandsSearch for more papers by this authorMichael Winklhofer, Michael Winklhofer Department of Earth and Environmental Sciences, Ludwig-Maximilians University, Munich, GermanySearch for more papers by this author Liao Chang, Corresponding Author Liao Chang Paleomagnetic Laboratory “Fort Hoofddijk,” Department of Earth Sciences, Utrecht University, Utrecht, Netherlands Research School of Earth Sciences, Australian National University, Canberra, Australian Capital Territory, AustraliaCorrespondence to: L. Chang, [email protected]Search for more papers by this authorIuliana Vasiliev, Iuliana Vasiliev Paleomagnetic Laboratory “Fort Hoofddijk,” Department of Earth Sciences, Utrecht University, Utrecht, NetherlandsSearch for more papers by this authorChristiaan van Baak, Christiaan van Baak Paleomagnetic Laboratory “Fort Hoofddijk,” Department of Earth Sciences, Utrecht University, Utrecht, NetherlandsSearch for more papers by this authorWout Krijgsman, Wout Krijgsman Paleomagnetic Laboratory “Fort Hoofddijk,” Department of Earth Sciences, Utrecht University, Utrecht, NetherlandsSearch for more papers by this authorMark J. Dekkers, Mark J. Dekkers Paleomagnetic Laboratory “Fort Hoofddijk,” Department of Earth Sciences, Utrecht University, Utrecht, NetherlandsSearch for more papers by this authorAndrew P. Roberts, Andrew P. Roberts Research School of Earth Sciences, Australian National University, Canberra, Australian Capital Territory, AustraliaSearch for more papers by this authorJohn D. Fitz Gerald, John D. Fitz Gerald Research School of Earth Sciences, Australian National University, Canberra, Australian Capital Territory, AustraliaSearch for more papers by this authorAnnelies van Hoesel, Annelies van Hoesel Department of Earth Sciences, Utrecht University, Utrecht, NetherlandsSearch for more papers by this authorMichael Winklhofer, Michael Winklhofer Department of Earth and Environmental Sciences, Ludwig-Maximilians University, Munich, GermanySearch for more papers by this author First published: 16 August 2014 https://doi.org/10.1002/2014GC005411Citations: 55AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Abstract Greigite (Fe3S4) is a widespread authigenic magnetic mineral in anoxic sediments and is also commonly biosynthesized by magnetotactic bacteria in aqueous environments. While the presence of fossilized bacterial magnetite (Fe3O4) has now been widely demonstrated, the preservation of greigite magnetofossils in the geological record is only poorly constrained. Here we investigate Mio-Pliocene sediments of the former Black Sea to test whether we can detect greigite magnetofossils and to unravel potential environmental controls on greigite formation. Our magnetic analyses and transmission electron microscope (TEM) observations indicate the presence of both diagenetic and bacterial greigite, and suggest a potentially widespread preservation of greigite magnetofossils in ancient sediments, which has important implications for assessing the reliability of paleomagnetic records carried by greigite. TEM-based chemical and structural analyses also indicate the common presence of nickel-substituted diagenetic iron sulfide crystals with a ferrimagnetic greigite structure. In addition, our cyclostratigraphic framework allows correlation of magnetic properties of Messinian Black Sea sediments (Taman Peninsula, Russia) to global climate records. Diagenetic greigite enhancements appear to be climatically controlled, with greigite mainly occurring in warm/wet periods. Diagenetic greigite formation can be explained by variations in terrigenous inputs and dissolved pore water sulfate concentrations in different sedimentary environments. Our analysis demonstrates the usefulness of greigite for studying long-term climate variability in anoxic environments. Key Points We provide evidence for the presence of biogenic greigite in ancient sediments Diagenetic greigite enhancements are climatically controlled Greigite is a paleoenvironmental indicator in anoxic environments 1 Introduction Greigite (Fe3S4) is a widespread iron sulfide mineral that has been found in anoxic sedimentary environments across the world over geologically significant time periods [Roberts et al., 2011a]. It is a strongly ferrimagnetic mineral that makes important contributions to paleomagnetic and environmental records. Paleomagnetic records carried by greigite can be challenging to interpret because greigite can form at different periods significantly later than deposition, which then leads to anomalous paleomagnetic records and remagnetization [e.g., Jiang et al., 2001; Roberts and Weaver, 2005; Rowan and Roberts, 2006; Roberts et al., 2010; Sagnotti et al., 2010]. But greigite can also form during earliest burial and, thus, record an excellent paleomagnetic signal [e.g., Tric et al., 1991; Vasiliev et al., 2004, 2005; Hüsing et al., 2007] that matches well with the geomagnetic polarity time scale (GPTS) [Cande and Kent, 1995] and provides crucial age control for sedimentary sequences. Greigite is also useful for paleoenvironmental analysis. For example, greigite abundances from the Santa Barbara Basin have been demonstrated to reflect millennial-scale climate variability [Blanchet et al., 2009]. In sulfate-reducing sedimentary environments, greigite often forms as a precursor to pyrite (FeS2). As an intermediate phase that forms during pyritization, greigite is expected to fully convert to pyrite during early diagenetic sedimentary sulfate reduction [Berner, 1984]. Kao et al. [2004] demonstrated that when reactive iron is abundant and dissolved sulfide concentrations are low in sedimentary pore waters, pyritization can be arrested and greigite can be preserved, because dissolved sulfide (H2S, HS–) is fully consumed by reactive iron. Roberts and Weaver [2005] argued that greigite can grow at any time during diagenesis when iron and sulfide are available. Formation and preservation of diagenetic greigite in sediments can be complex and can preclude straightforward interpretation of magnetic signals carried by greigite. Greigite is also often produced by sulfidic magnetotactic bacteria (MTB) in natural environments [Farina et al., 1990; Mann et al., 1990]. MTB intracellularly biomineralize chains of greigite or magnetite (Fe3O4) crystals as microscopic compasses that enable them to move along geomagnetic field lines to find optimal living conditions [Bazylinski and Frankel, 2004]. The inorganic remains of MTB preserved in sediments as magnetofossils can make significant contributions to paleomagnetic signals [e.g., Kirschvink and Chang, 1984; Petersen et al., 1986; Stoltz et al., 1986; Roberts et al., 2012; Heslop et al., 2013] and can reflect environmental processes, such as glacial-interglacial fluctuations [Hesse, 1994; Yamazaki, 2012; Heslop et al., 2013], hyperthermal events [Kopp et al., 2007; Schumann et al., 2008; Chang et al., 2012; Larrasoaña et al., 2012], and oceanic productivity [Roberts et al., 2011b; Yamazaki and Ikehara, 2012; Chang et al., 2013]. Unlike diagenetic greigite, which can remagnetize sediments or overprint primary remanences due to postdepositional growth, bacterial greigite should carry a syndepositional paleomagnetic signal that is useful for magnetostratigraphic dating and paleomagnetic reconstructions [Vasiliev et al., 2008]. Compared to magnetite-producing MTB, greigite producers are less common in modern environments, although they have been identified in many sulfide-rich aquatic environments [e.g., Farina et al., 1990; Mann et al., 1990; Bazylinski et al., 1995; Pósfai et al., 1998a, 1998b; Simmons et al., 2004; Reitner et al., 2005; Wenter et al., 2009; Lefèvre et al., 2011; Wang et al., 2013]. Preservation of magnetite magnetofossils in sediments was expected to be common since the discovery of MTB in the 1970s [Kirschvink and Chang, 1984; Petersen et al., 1986; Stoltz et al., 1986; Hesse, 1994; Hounslow and Maher, 1996], and they are now being frequently reported [Kopp et al., 2007; Kopp and Kirschvink, 2008; Roberts et al., 2011b, 2012; Chang et al., 2012; Larrasoaña et al., 2012; Yamazaki, 2012; Heslop et al., 2013]. In contrast, geological preservation of greigite magnetofossils is not well documented. Putative greigite magnetofossils have been reported from Miocene and Pliocene sediments from the Carpathian foredeep, Romania [Pósfai et al., 2001; Vasiliev et al., 2008]. Bacterial greigite has also been identified within Baltic Sea sapropels up to several thousand years in age [Reinholdsson et al., 2013]. Similar magnetic properties suggest that biogenic greigite may also magnetically dominate some Mediterranean sapropels [Roberts et al., 1999; Reinholdsson et al., 2013]. In order to assess the preservation potential of greigite magnetofossils in the geological record and to study greigite formation under changing environmental and climatic conditions, we use rock magnetic techniques, transmission electron microscope (TEM) observations, X-ray fluorescence (XRF), and cyclostratigraphy, to investigate Mio-Pliocene Black Sea sediments. 2 Geological Setting and Astronomical Tuning Samples from several Mio-Pliocene sedimentary sections in the former Black Sea basin in Russia and Romania were selected for this study (Figure 1). Samples labeled “TK-XXX” and “TR-XXX” are from the Zheleznyi Rog (Iron Cape) section on the Black Sea margin of the Taman Peninsula, Russia. Standard paleomagnetic drill cores were taken at ∼1–2 m stratigraphic intervals. This section is 500 m thick and covers ∼5 Myr of Mio-Pliocene sedimentation in the Black Sea domain of the Paratethys Sea [Vasiliev et al., 2011]. We focus on the upper 250 m of the section (Pontian and Kimmerian regional stages) [Krijgsman et al., 2010]. This is the time equivalent of the Messinian Salinity Crisis (MSC; 5.97–5.33 Ma) [Manzi et al., 2013; Roveri et al., 2014]. The base of the Pontian is characterized by a marine flooding event, which represents a reconnection of the Black Sea to the Mediterranean Sea [Krijgsman et al., 2010]. This event was followed by reestablishment of brackish conditions in the Black Sea basin and a major invasion of Pannonian species [Stoica et al., 2013; Grothe et al., 2014]. Figure 1Open in figure viewerPowerPoint Map of the Black Sea domain of the Paratethys region with locations of the studied sedimentary sections in the Taman Peninsula, Russia (red star) [Krijgsman et al., 2010] and in Romania (white stars) [Vasiliev et al., 2004, 2005], where putative greigite magnetofossils have been reported [Vasiliev et al., 2008]. The Pontian sedimentary succession is characterized by distinct dark-light alternations (labeled A–D and 1–11 in Figure 2a), which indicate stable depositional conditions and a potential link to (astronomically) forced climate cyclicity. The alternations pass upward from dark bedded bituminous clay and marl, to diatomite couplets (A–C), to light and dark gray marls (D, 1–11). The diatomite couplets reflect increased marine conditions compared to the interbedded marls and represent diatom blooms at the transition between marine and freshwater conditions [Radionova and Golovina, 2011; Radionova et al., 2012]. Straightforward correlation to either obliquity or precession cannot be made; correlation to the astronomically tuned δ18O record of Van der Laan et al. [2005, 2006] provides a more straightforward pattern fit. This record from the Atlantic site of Morocco is the nearest oxygen isotope record not influenced by MSC evaporites and is taken as representative of regional climate. In our interpretation, light intervals correlate to cold/glacial stages and darker intervals to warmer/interglacial stages. The darker marls may represent a higher terrestrial component, related to higher river discharge during interglacial periods. In glacial periods, this is reduced and the “marine” calcite component is more dominant. Figure 2Open in figure viewerPowerPoint Chronology and climate stratigraphy for the Zheleznyi Rog section on the Black Sea coast, Taman Peninsula, Russia. (a) Field photograph with clear dark-light alternations related to climatic cyclicity. (b) Tentative astronomical tuning of the cyclic sediments to the La2004 astronomical solution [Laskar et al., 2004] and benthic δ18O records for the upper Miocene Ain el Beida section, northwestern Morocco [Van der Laan et al., 2005, 2006]. Numbers in Figure 2a correlate to the numbers in the lithostratigraphic log in Figure 2b. The green circle in (a) indicates people in the field as a scale. The geomagnetic reversal at 180 m is an age correlation tie-point that corresponds to the C3An.1n-C3r reversal at 6.033 Ma [Krijgsman et al., 2010; Vasiliev et al., 2011]. Our cyclostratigraphic correlation suggests a flooding age for the base of the Pontian to have been around 6.1 Ma, rather than the previously suggested 6.04 Ma [Krijgsman et al., 2010]. The majority of the Pontian was deposited during the long C3r reversed chron, which provides no additional magnetostratigraphic tie-points. The pattern of light beds 3–8 is, however, distinct and appears to correlate well with glacial δ18O stages TG 24 to TG 20 (Figure 2b). Overlying the Pontian marls is a distinct reddish interval that contains oolites/pisolites and nodular marls, which suggest high-energy coastal deposition. This reddish layer marks the base of the Kimmerian stage and is interpreted as the MSC climax related to glacial peaks TG 12 and TG 14, which recorded a sea level drop in the Black Sea basin [Krijgsman et al., 2010]. Faunal indicators suggest a change from brackish-marine conditions in the Pontian stages to a freshwater environment in the Kimmerian [Krijgsman et al., 2010], where no clear lithological cyclicity is observed. Selected samples (“BD-XXX” and “RR-XXX”) from Pliocene Black Sea sediments of the Romanian Carpathians [Vasiliev et al., 2004, 2005] were also analyzed. These sediments, consisting of blue to gray sandstones, siltstones and clays, were deposited at high sedimentation rates (60–150 cm/kyr) in brackish environments. The samples are the same as those studied by Vasiliev et al. [2008]. Several synthetic greigite samples (“S-XXX”) [Chang et al., 2007, 2008] and magnetite magnetofossil-bearing sediment samples were analyzed to compare with results from greigite-bearing sediments. Sample “CD143-1-21” is a surface sediment sample (21 cm below the surface) from piston core CD143–55705 from the Oman margin [Rowan et al., 2009]. Sample “738C-11R1–40” is a pelagic carbonate from the Paleocene-Eocene thermal maximum from ODP Hole 738C, Southern Ocean [Larrasoaña et al., 2012]. Sample “689D-11H6–21” is an early Oligocene pelagic carbonate from ODP Hole 689D, Southern Ocean [Roberts et al., 2012]. 3 Methods Room temperature hysteresis, isothermal remanent magnetization (IRM) acquisition, backfield demagnetization, and FORC measurements were performed using a Princeton Measurements Corporation MicroMag alternating gradient magnetometer (AGM) (Model 2900; noise level 2 × 10−9 Am2) at the Paleomagnetic Laboratory of Utrecht University. Hysteresis loops were measured between ±1 T with a field step of 4 mT and a 150 ms averaging time. Hysteresis parameters, including the saturation magnetization (Ms), the saturation remanent magnetization (Mrs), and coercivity (Bc), were determined after paramagnetic slope correction. FORC diagrams [Pike et al., 1999; Roberts et al., 2000] were obtained with maximum applied fields of 1 T, field increments up to 0.4 mT, and averaging times of 100–300 ms. FORC diagrams were calculated using a program written by Tom Mullender at the Paleomagnetic Laboratory, Utrecht University. Data representation close to the vertical axis of the FORC diagrams may be biased due to data extrapolation in this region. To avoid uncertain interpretation in this poorly defined part of the FORC space, it is not considered by the software. A smoothing factor (SF) between 3 and 5 was used. IRM acquisition curves were obtained by measuring 100–150 logarithmically spaced field steps up to a maximum field of 1 T on the AGM. IRM curves for some magnetically weak samples were measured using standard paleomagnetic core samples (∼2 × 2.5 cm) on a robotized superconducting rock magnetometer (noise level 1–2 × 10−12 Am2) at Utrecht University. Sixty data points were measured up to 700 mT. IRM acquisition curves were decomposed into coercivity components using the fitting protocol of Kruiver et al. [2001]. Thermomagnetic runs were measured in air with a modified horizontal translation-type Curie balance with a sensitivity of ∼5 × 10−9 Am2 [Mullender et al., 1993] at Utrecht University. The applied field was cycled between 100 and 300 mT. Multiple heating and cooling cycles between room temperature, 200, 300, 350, 450, 620, and 700°C were performed at a heating/cooling rate of 6°C/min. Depending on sample magnetization, ∼10–100 mg of sediment was used. A quartz glass holder and quartz wool were used to hold samples. Low-temperature magnetic properties were measured with a Quantum Design Magnetic Property Measurement System (MPMS; model XL7) at the Australian National University (ANU). For zero-field cooled (ZFC) and field-cooled (FC) curves, samples were cooled to 10 K in either zero-field or a 5 T field. At 10 K, a 5 T field was applied and then switched off to impart a low-temperature saturation IRM (SIRM), and the MPMS magnet was reset (the residual field after a magnet reset is ∼200–300 μT). SIRM warming curves were measured during warming in zero-field. ZFC warming curves were acquired before the FC curves. Ferromagnetic resonance (FMR) spectroscopy is used widely to detect biogenic magnetite chain signatures [e.g., Weiss et al., 2004; Kopp et al., 2006, 2007; Roberts et al., 2011b, 2012; Chang et al., 2014] and was employed to study our greigite-bearing samples. X-band FMR spectra were measured with a JEOL electron paramagnetic resonance (EPR) spectrometer at the Technical University of Munich or a Bruker Elexys EPR spectrometer at ANU. The 30–100 mg of sediment was measured at a frequency of 9.1–9.4 GHz and power of 0.6–2 mW. We used the same FMR parameters as those in other studies [Kopp et al., 2006; Roberts et al., 2011b, 2012]. Magnetic particles were extracted from bulk sediments by adapting the methods of Petersen et al. [1986] and Vasiliev et al. [2008]. Approximately 10 g of ground sediment was mixed with 300 mL of argon-purged demineralized water. Sodium polyphosphate (Na4P2O7·10H2O) was added to disperse the clays. The sediment solution was agitated using an ultrasonic bath for 30 min. The obtained sediment suspension was circulated through a Frantz isodynamic magnetic separator for several hours. Magnetic extracts were then washed several times with demineralized water using a bar magnet. Magnetic extracts were viewed and analyzed using either a FEI Tecnai 20 FEG TEM at an acceleration voltage of 200 kV at Utrecht University, or a Philips CM300 TEM operated at 300 kV at ANU. Elemental compositions and crystal structures were analyzed using energy dispersive spectroscopy (EDS) and selected-area electron diffraction (SAED). XRF analysis was carried out using a Thermo Scientific Niton XL3t XRF Analyzer at Utrecht University. Measurements were made on the top surface of standard paleomagnetic cylindrical samples. One flat surface of the cylinder was cleaned directly prior to data acquisition to ensure that data were obtained from fresh samples. Three measurements were averaged. Data that were significantly different from other measurements (e.g., when the X-ray beam interacted with large fossil shells) were excluded and measurements were repeated. Our XRF analysis is semiquantitative because no standard samples were used for calibration. Our purpose is to track elemental trends through the sequence, rather than absolute elemental concentrations. There are good correlations between large-scale elemental variations and magnetic properties, which are expected to reflect real sedimentary variations. Samples from the Zheleznyi Rog section were taken at ∼1–2 m stratigraphic intervals for paleomagnetic analysis [Krijgsman et al., 2010; Vasiliev et al., 2011]. Sister paleomagnetic samples were used for rock magnetic, XRF, and electron microscopic analyses. A total of ∼150 samples from Taman (selected from the 0–230 m stratigraphic interval) were measured for hysteresis properties, IRM acquisition and backfield curves, and XRF elemental variations; ∼40 were measured for FORC diagrams, ∼20 were selected for thermomagnetic analyses, and 6 were selected for low-temperature magnetic analyses, FMR, and TEM observation. For samples from Romania, ∼20 paleomagnetic drill core sister samples [Vasiliev et al., 2004, 2005, 2008] were selected for hysteresis, IRM, FORC, FMR, and thermomagnetic analyses. 4 Results Rock Magnetism 4.1.1 Hysteresis Properties and FORC Diagrams The studied samples have a wide range of hysteresis properties, mostly with single domain (SD) behavior (Figures 3a–3c). One type of hysteresis loop has relatively square shapes with Mrs/Ms ratios of ∼0.28–0.43 and Bc values of ∼10–25 mT (Figure 3a). The other type has high Bc values (mostly between 40 and 50 mT) and high Mrs/Ms ratios up to ∼0.7, which are dominated by cubic magnetocrystalline anisotropy [Roberts, 1995] (Figure 3b). Some samples have intermediate Bc and Mrs/Ms values mostly between these two types of hysteresis loops (Figure 3c). We classified the studied SD-dominated samples into three groups by their hysteresis and FORC characteristics: type-A (lower Bc and square loops), type-B (higher Bc), and type-C (intermediate between type-A and type-B). We use this classification consistently throughout this paper. Hysteresis ratios for some samples are plotted after Day et al. [1977] in supporting information Figure S1. Figure 3Open in figure viewerPowerPoint Comparison of rock magnetic results for three types of studied greigite samples and a biogenic magnetite sample. (a–c) Hysteresis loops, (d–f) FORC diagrams, (g–i) IRM component analyses, (j–l) thermomagnetic curves, (m–o) low-temperature ZFC-FC curves, and (p–r) FMR spectra. A fourth column contains magnetic data for typical magnetite magnetofossil-bearing sediment (sample “738C-11R1–40”) from the Palaeocene-Eocene thermal maximum, ODP Hole 738C [Larrasoaña et al., 2012]. All hysteresis loops (Figures 3a–3c) are normalized to Ms values (at 1 T) after paramagnetic slope correction. Hysteresis parameters (Bc and Mrs/Ms), determined from the entire loop to ±1 T, are indicated. FORC diagrams for type-A samples with a narrow central ridge along the Bi axis, with negligible vertical spread, are indicative of negligible magnetostatic interactions. FORC diagrams for type-B samples have concentric contours and a large vertical spread. Type-C samples have mixed FORC signatures from both noninteracting and interacting components. For results of IRM component analyses (Figures 3g–3i), squares indicate measured data. Red lines are the fitted total spectra. The other lines (magenta, green, and blue) are different components with lognormal coercivity distributions. Thermomagnetic runs (Figures 3j–3l) were measured in air during multiple heating and cooling cycles (arrows). All samples produce an irreversible magnetization decrease below 400°C, due to greigite decomposition. The peak at ∼500°C is due to oxidation of pyrite into magnetite [Passier et al., 2001]. In the low-temperature magnetic data (Figures 3m–3o), FC and ZFC SIRM warming curves are indicated by red solid circles and blue open circles, respectively. The magnetizations are normalized to the FC data at 10 K. No clear Verwey transition is observed. In the FMR spectra (Figures 3p–3r), arrows indicate paramagnetic high-spin Fe3+ signals (g = 4.3) and Mn2+ sextet signals. Raw data are shown in red. Black lines are smoothed curves that reduce the effects of the sharp Mn2+ EPR sextet signals and noise. FMR parameters for each measured sample are shown. High-resolution FORC measurements indicate variable coercivity and magnetostatic interaction distributions (Figures 3d–3f). Type-A samples (Figure 3d) have a dominant narrow ridge along the Bi axis with negligible vertical spread. This FORC signature corresponds to that of noninteracting or weakly interacting SD particles [Pike et al., 1999; Egli et al., 2010]. Type-B samples have FORC diagrams (Figure 3e) with concentric contours with large vertical spread indicative of strong magnetostatic interactions [Pike et al., 1999; Roberts et al., 2000]. FORC diagrams for type-C samples have a mixture of type-A and type-B signatures: with both a central ridge and vertically spread concentric contours (Figure 3f). These samples, therefore, contain a variable mixture of noninteracting and interacting SD particles. FORC measurements for the same type-C samples measured with different numbers of FORCs but covering the same FORC region indicate that the central ridge signature is not clear in low-resolution FORC diagrams (Figures S2a and S2b), but it becomes more clearly evident in high-resolution FORC diagrams (Figures S2c and S2d). 4.1.2 IRM Acquisition and Demagnetization Coercivity of remanence (Bcr) values, determined from backfield demagnetization measurements, range between ∼30 and 70 mT. IRM acquisition curves for most samples can be fitted [Kruiver et al., 2001] with two main components (Figures 3g–3i). For type-A samples, one IRM component has a median field of ∼50–70 mT and a small fitted dispersion parameter (DP) in the range of 0.20–0.22 (Figure 3g, magenta line). This component contributes ∼55–80% of the total IRM. The other IRM component has a smaller median field (∼25–36 mT) and broader distribution (DP = ∼0.45–0.48) (Figure 3g, green line). Type-B samples also have two major IRM components, one with a narrow distribution (Figure 3h, magenta line) and the other with a broad distribution (Figure 3h, green). The narrow IRM component for type-B samples contributes ∼60–70% of the total IRM intensity. Compared to type-A samples, the narrow IRM component for type-B samples has higher median field (mostly between ∼75 and 80 mT). DP values for this component are even smaller (mostly between 0.13 and 0.16). The other major component with broader distribution for type-B samples has median field values of ∼40–67 mT, and DP values of ∼0.32–0.36. Type-C samples (Figure 3i) have IRM components similar to type-B samples. All types of samples have an IRM component with DP < 0.22. The small low-field IRM component for all types of samples likely reflects deviations of the major IRM component from an ideal lognormal model distribution [Egli, 2004], to which we do not attribute physical meaning. This “skewed-to-the-left” behavior requires an extra component if fitting is restricted to symmetric distributions [Kruiver et al., 2001]. 4.1.3 High and Low-Temperature Magnetic Properties Thermomagnetic runs in air during different heating and cooling cycles indicate an irreversible decrease in magnetization between 200 and 400°C with most pronounced drops in magnetization from ∼350 to 420°C in many samples (Figures 3j–3l). An irreversible decrease in magnetization between 200 and 420°C is indicative of greigite [e.g., Reynolds et al., 1994; Roberts, 1995; Dekkers et al., 2000; Vasiliev et al., 2008; Chang et al., 2008; Roberts et al., 2011a] and is not characteristic of stoichiometric magnetite. For oxidized magnetite, such as magnetite with a maghemite (γ-Fe2O3) coating, further oxidation occurs during heating in air, which causes irreversible magnetizatio