2016 Volume 23 Pages 41-51
In order to determine magnetic mineral distributions in nearshore marine sediments off southwestern Chile, chemical analysis of bottom waters and rock magnetic characterization of surface sediments were performed. The samples analyzed are terrigenous and calcareous sediments recovered with a multiple corer at five stations. Calcareous sediments occur at stations with highly dissolved oxygen catchment in the bottom waters. Concentration-dependent magnetic parameters of calcareous sediment samples show relatively low values, and their magnetic grain sizes are coarse. Higher values of concentration-dependent magnetic parameters are recognized, and fine grain magnetic minerals are distributed in silty and sandy clay sediments. Magnetic mineralogy of the sediment samples reflects not only lithology but also redox conditions in the overlying bottom waters. Thermo-magnetometry results suggest that magnetite (Fe3O4), hematite (αFe2O3), (titano)maghemite (γFe2O3), and goethite (αFeOOH) are common in sediment samples at all stations. Higher coercivities and total organic carbon (TOC) values are also observed at oxic stations, suggesting the presence of goethite and (titano)maghemite. In general, higher TOC contents in sediments lead anoxic conditions due to organic matter decomposition, therefore goethite and (titano)maghemite are unstable. Despite higher TOC contents in this sediment sample, such magnetic minerals can present. This implies that the magnetic minerals are kept in oxic conditions.
Rock magnetic properties of nearshore marine sediments have been used as proxies of detrital mineral supply changes (e.g., Bloemendal et al., 1992; Walden et al., 1999; Zhang et al., 2008; Liu et al., 2012). Nearshore marine sediments generally have high sedimentation rates because of larger terrigenous sediment inputs, and are expected to preserve as high-resolution records of paleo-environmental and paleoclimate variations (e.g., Bloemendal et al., 1992; Evans and Heller, 2003; Maher 2011; Liu et al., 2012). However, magnetic minerals can dissolve and/or form, depending on redox conditions within sediments (e.g., Berner 1980; Karlin and Levi, 1983, 1985, Henshaw and Merrill, 1980; Canfield and Berner, 1987; Leslie et al., 1990a, b; Karlin, 1990; Roberts and Turner, 1993; Liu et al., 2004; Garming et al., 2005; Rey et al., 2005; Kawamura et al., 2008; 2012; Rowan et al., 2009; Liu et al., 2012). In order to interpret paleomagnetic records from nearshore sediments, understanding of magnetic mineral assemblages is required, and what are reflected and recorded to them is essential.
The studied nearshore sediments are from the southeastern Pacific Ocean along the southwestern part of Chile, which is characterized by a complex system of fjords and channels (Figure 1). Strong poleward winds cause heavy precipitation. Ice melting and rainfall greatly enhance the supply of terrigenous sediments to the area. Fresh water is lighter than saltwater and remains at sea surface, which causes surface salt water to sink. As a result, the Subantarctic Mode Waters (the precursor of the Antarctic Intermediate Water) forms along the coast of southern Chile, which has a large impact on ocean circulation (e.g., Warren and Wunsch, 1981; Downes et al., 2009). Nearshore sediments in this study area are generally rapidly deposited, and are expected to be good archives of paleo-environmental changes. The aim of this study is to understand how magnetic minerals assemblages and the grain sizes of magnetic minerals are affected by the redox state of overlying bottom waters along the southwestern coast of Chile. This study reports the analytical results of bottom water and sediments, which are taken from various redox states at five stations.
(a) Map of the western hemisphere with the study area indicated (b) detailed map of the study area. Solid squares show the stations in this study.
Samples were collected at five stations in the south Pacific bays offshore southwestern Chile during the MR08-06 cruise of the R/V Mirai (Japan Agency for Marine-Earth Science and Technology) in March, 2009 (Figure 1; Table 1). A multiple corer (Rigo Co., Ltd.) was used to sample the sediment-water interface. The recovered sediments consist of terrigenous and calcareous silty clay, sandy clay, and/or sand (Table 1; Harada, 2008).
Sampling locations, lithology, total nitrogen (TN), organic carbon (TOC), sulfur (TN) of the sediments, and bottom water profiles.
Dissolved oxygen (DO) of bottom waters was measured directly with a DO meter (Horiba Co., Ltd., OM-51-2) onboard immediately after recovering of the multiple cores. Dissolved iron in bottom waters was prepared for analysis according to Achterberg et al. (2001). Bottom water just above the sediment-water interface was taken from the cores using a plastic syringe. The water samples were passed through a filter (pore diameter of 0.45 $\mu$m), and the samples 30 mL were stored in Teflon bottles. The water samples were treated with 1 mL of special grade nitric acid (1 mol/L concentration), and pH was kept below 1 at room temperature. The concentration of dissolved iron (DI) was measured with a flameless graphite furnace atomic adsorption spectrometer (Shimadzu Co., Ltd., AA-6800) at the Japan Coast Guard Academy (JCGA). The samples were diluted 6-10 times for the measurements. Standard samples were made by diluting an iron standard solution (Nakarai Tesque Co., Ltd.) to 0.05-0.1 nmol/L using pure water, and by adding 400 $\mu$L of special grade nitric acid. DI measurements were carried out seven times per sample, and the averages were calculated.
3.2 Total organic carbon (TOC) and total nitrogen (TN) in the sedimentsConcentrations of total organic carbon (TOC) and total nitrogen (TN) in the sediments were analyzed using a CHNS (carbon, hydrogen, nitrogen, and sulfur) analyzer (vario EL III, Elementar Co. Ltd.) at JCGA. Dried and powdered sediment samples of approximately 20 mg were used. TOC/TN ratios represent the sources of organic materials in the sediments, and its high value indicates that the sediments are rich in organic materials of detrital origin (e.g., Emerson and Hedges, 1988).
3.3 Rock magnetic analysesWet samples were taken from the top of five multiple cores using 6.7 cm3 plastic cubic boxes (Natsuhara Giken). The cube samples were used to measure the low-field magnetic susceptibility $(\chi )$, natural remanent magnetization (NRM), anhysteretic remanent magnetization (ARM), and isothermal remanent magnetization (IRM) at Kyoto University. $\chi $ was measured with a Kappabridge KLY-3S magnetic susceptibility meter. NRMs and ARMs were measured with a 2-G Enterprises model 760 superconducting rock magnetometer. ARMs were imparted to samples using a steady direct current (DC) bias field of 0.1 mT in a peak alternating field (AF) of 100 mT. $\chi$ARM was calculated by dividing the ARM by the value of the steady bias field (Banerjee et al., 1981). A saturation IRM (SIRM) was imparted in a 2.5 T inducing field and an IRM was induced with a backfield of 0.3 T along one sample axis. These IRMs were measured with a spinner magnetometer (Natsuhara-Giken SMM-85). The hard IRM (HIRM) and S－0.3 T were calculated according to the definitions of Bloemendal et al. (1992):
| \[ \text{HIRM} = (\text{SIRM} + \text{IRM}_{-0.3\, \text{T}})/2,\ \text{and} \] \[ \text{S}_{-0.3\, \text{T}} = (1 - \text{IRM}_{-0.3\, \text{T}}/\text{SIRM})/2. \] |
All cube samples were dried in an oven at 40℃ after the measurements. Each mass-specific rock magnetic parameter was calculated using the weight of the dried cubic sample.
In order to determine magnetic minerals in the samples, thermal demagnetization of composite IRMs (Lowrie, 1990) were conducted. The dried powder samples (ca 50 mg) were packed in a small quartz cup (5 mm in diameter and 10 mm in height). A magnetic field of 2.5 T was applied along the vertical direction of the cup, and then fields of 0.3 T and 0.07 T were applied successively along the two remaining perpendicular axes using a pulse magnetizer (Magnetic Measurements Ltd. model MMPM-9). A magnetic field of 0.07 T was used to detect magnetite and greigite (e.g., Oda and Torii, 2004). The remanent magnetization was measured with a 2-G Enterprises 760R magnetometer at Kyoto University. Low-temperature magnetometry was conducted using a Quantum Designs magnetic property measurement system (MPMS-XL5) at the Center for Advanced Marine Core Research, Kochi University. Approximately 50 mg chips were taken from the dried samples and were packed in a capsule. An IRM was imparted to the samples at 5 K in a 2.5 T field, and the remanent intensities were measured from 5 K to 300 K in a zero-field at 2 K intervals.
Hysteresis loop measurements and direct-current demagnetization (DCD) of a saturation remanent magnetization (Mrs) were performed on 10-30 mg sediment chips with an alternating gradient magnetometer (AGM, Model 2900-02, Princeton Measurements Corporation) at Kyoto University. Five sediment chips per station were measured. The maximum applied field was 1.0 T and the field increment was 2 mT during hysteresis and DCD measurements. Coercivity (Hc), saturation magnetization (Ms), and saturation remanent magnetization (Mrs) were determined from the hysteresis measurements. The coercivity of remanence (Hcr) was determined from DCD results.
First-order reversal curve (FORC) measurements were conducted using an alternating gradient magnetometer (AGM, Princeton MicroMag 2900). The field spacing between measurements was set to 1 mT. A total of 193 FORCs were measured, with Hc between 0 mT and 80 mT, and Hu between －50 mT and 50 mT. The software ''FORCinel ver. 2.03'' of Richard Harrison (https://wserv4.esc.cam.ac.uk/nanopaleomag/) was used for data processing, and a smoothing factor of 3 was adopted (Harrison and Feinberg, 2008).
3.4 Principal component analysisIn order to clarify the relationship between redox conditions in bottom waters and magnetic properties of the sediment samples, principal component analysis was performed on a matrix of correlation coefficients among values of DO, DI and rock magnetic parameters shown in Figure 2. The software ''Excel Statistics for Windows'' (Social Survey Research Information Co., Ltd.) was used. Results are listed in Table 2.
Dissolved oxygen (DO) and vertical distribution of dissolved iron (DI) with error bars in the bottom waters, and magnetic parameters of surface sediments, including magnetic susceptibility $(\chi)$, susceptibility of anhysteretic remanent magnetization ($\chi$ARM), natural remanent magnetization (NRM), saturation isothermal remanent magnetization (SIRM), hard IRM (HIRM), S-ratio (S－0.3 T), total nitrogen (TN), and total organic carbon (TOC). DI average values are shown as solid circles with error bars.
Principal component analysis of rock magnetic parameters of the sediments and DO and DI in bottom water.
DI and DO values of bottom waters are shown with water depths of the five stations in Figure 2. DO values are relatively low at stations 38 and 43, while high values are found at stations 44, 45, and 46. The physical structure of the water column in this area was investigated using a conductivity temperature depth profiler (Harada, 2008), and vertical profiles of DO values were measured at all stations in this study. DO values of bottom waters at the five stations of Harada (2008) are coincident with our data. DI averages did not vary significantly among the stations. However, variations in DI are similar to DO. This means that the concentration of iron hydroxides in suspended solids (<0.45 $\mu$m in diameter) are highly to be present in the relatively oxic bottom waters (e.g., Aston and Chester, 1973).
4.2 TN and TOC in the sediment samplesTN and TOC values are listed in Table 1 and Figure 2. The lowest values of TN and TOC are recognized at station 43. TOC values are low at stations 38 and 43, which consists of silty clay and sandy clay with nannofossils and are under relatively anoxic bottom waters. TOC contents of the sediment samples from stations 44, 45, and 46 indicate relative high values, and are consisted of foraminiferal and/or calcareous clayey sand. TOC/TN ratio shows high values at stations 44 and 45. It means that terrestrial organic matters are abundant in the sediment samples.
4.3 Rock magnetic parametersRock magnetic parameters are shown in Figure 2. The concentration-dependent rock magnetic parameters, such as $\chi $, $\chi$ARM, and SIRM, have the highest values at station 43. A sample from station 38 shows the second highest values except for $\chi$ARM. Samples from the other stations (44, 45 and 46) have relatively low values, which consist of calcareous sands with foraminifera. NRM values show a similar pattern to SIRM values. HIRM variation resembles those of $\chi $ and SIRM; higher HIRM values occur at stations 38 and 43. S－0.3 T variation is similar to those of the concentration-dependent magnetic parameters. S－0.3 T is higher at stations 38 and 43 than the other stations, which suggests that the proportion of low coercivity magnetic minerals is high at these stations. The minimum value of S－0.3 T is observed at station 44 with high DO and DI values in bottom waters, whereas samples from stations 38 and 43 with relatively lower DO and DI values in bottom waters show relatively higher S－0.3 T values.
Thermal demagnetization results of composite IRMs for samples from all stations are shown in Figure 3. Soft (<0.07 T), and medium (0.07-0.3 T) components are demagnetized completely at around 580℃, which is the Curie point of magnetite, for samples from stations 38, 44, 45, and 46 (Figure 3). Slight thermal decay of the hard components (<2.5 T) is observed at around 675℃, which is the Curie point of hematite, in all samples. An inflection in demagnetization curves at around 300℃ is recognized in samples from all stations (Figure 3). Authigenic greigite (e.g., Roberts et al., 2011) is not expected to form under an oxic water column. The inflection suggests the presence of (titano)maghemite (e.g., Özdemir and Banerjee, 1984; Ishikawa and Frost, 2002). A decrease at around 620℃ is shown in all samples (Figure 2), which also suggests the presence of maghemite (e.g., Heider et al., 1992). The remanent magnetization intensities of the hard component decreases at 120℃ in stations 38 and 43 samples (Figures 3a and 3b), which corresponds to the Néel temperature of goethite (e.g., Bocquet and Kennedy, 1992; Özdemir and Dunlop, 1996). The intensity of the high coercivity component is higher at station 44 than the other stations (Figure 3c).
Stepwise thermal demagnetization of composite three-axis isothermal remanent magnetization (IRM) (Lowrie, 1990) for surface sediment samples (0-2 cm depth). Coercivity spectra are shown with respect to temperature for the 0-0.07 T, 0.07-0.3 T, and 0.3-2.5 T components. The Néel temperature of goethite at 120℃, the thermal decomposition of (titano-)maghemite at around 300-400℃, the Curie point of magnetite at 580℃, titanium-poor maghemite at 620℃ (Heider et al., 1992), and the Néel temperature of hematite (Ht) at 675℃.
Results of low-temperature magnetometry are show in Figure 4. The IRM warming curves for all samples show clear decreases at around 120 K (Figure 4), which represents the Verwey transition of magnetite (Verwey, 1939). The decline from 5 K to 50 K in the station 44 sample suggests the presence of goethite (e.g., Liu et al., 2006) or it is attributable to paramagnetic minerals.
Thermal demagnetization curves for a low-temperature IRM imparted at 5 K with a 2.5 T DC field. Samples were analyzed from surface sediments (0-2 cm) at five stations. Solid squares represent IRM intensity. Open squares with solid squares correspond to dIRM/dT.
Magnetic grain size dependent parameters (Mrs/Ms and Hcr/Hc) are shown as a Day plot (Day et al., 1977) in Figure 5. Most samples from stations 38 and 43, which have low DO values of bottom water and consist of sandy or silty clays with nannofossils, are located in or around the pseudo-single domain (PSD) region (Figure 5). Data for samples from stations 44, 45, and 46, which consist of foraminiferal or calcareous clayey sands (Table 1), plot in or around the multi domain (MD) region of the Day plot. FORC diagrams of the sediment samples are also shown in Figure 5. A component that spread to Hu-axis direction and low Hc values, which is attributable to MD were recognized in all samples. Especially the MD component was dominant in station 44 sample. A ridge along the horizontal axis is recognized at stations 38, 43, and 46 samples (Figures 5b, 5c, and 5d), while the ridge is vague in stations 44 and 45 sediment samples (Figure 5e and 5f). It means that non-interacting SD magnetic minerals may be included in the samples in stations 38, 43, and 46 samples.
Day plot (cf. Day et al., 1977), hysteresis loops, and FORC diagrams of samples from the studied stations. Single domain (SD), pseudo-single domain (PSD), and multi-domain (MD) regions are indicated. See text for discussion
According to correlation coefficients and coefficients of determination (Table 2), DI shows higher positive relationship with DO, and Mrs/Ms, NRM, SIRM, HIRM, and S－0.3 T have relatively higher negative relationship with DO. The result of the principal component analysis indicates that eigenvalues of the first component with the proportion of 76.7％ are relatively high in DO with negative sign and in NRM, SIRM, HIRM and S－0.3 T with positive one (Tables 2-1 and 2-2). The scores of the first component for the five stations decrease with increasing DO values (Tables 1 and 2-3). It may be inferred that the first component is a factor representing the negative correlation between DO and the rock magnetic parameters with higher eigenvalue. The second principal component with the proportion of 13.2％ shows a higher positive eigenvalue in Mrs/Ms, and higher negative ones in Hcr/Hc, $\chi$, and $\chi $ARM. The meaning of the second component is obscure.
Variations in magnetic concentration-depended parameters seem to correspond to lithology. The lithology at the five stations is listed in Table 1. At stations 44 and 45 (foraminiferal sand) and station 46 (calcareous clayey sand), the concentration-dependent parameters ($\chi$, SIRM, and HIRM) have low values. At stations 38 and 43 (silty clay and sandy clay with nannofossils), these parameters have high values (Figure 2). The rough lithological difference is recognized among the stations. Calcareous grains are diamagnetic and has a weak negative response to an applied magnetic field (e.g., Thompson and Oldfield, 1986). Ferromagnetic minerals in the sediments at stations 44, 45, and 46 would be diluted by such diamagnetic materials and/or non-magnetic minerals, thus the concentration of ferromagnetic mineral is low.
Compositions of magnetic minerals in the sediment samples are locally different. Magnetite, maghemite, and hematite are common at the all stations (Figures 3 and 4). The decline of the hard components at around 80-120℃ in thermal demagnetization indicative of goethite is observed at stations 38 and 43. S－0.3 T values are relatively high at stations 38 and 43, which are also characterized by relatively low DO values in the bottom water. On the other hand, S－0.3 T of stations 44, 45, and 46 are low, where DO values are relatively high. Maghemite is known as an iron oxide distributed under an oxic environment, and has higher coercivity than magnetite (e.g., Walden et al., 1999; Laurent et al., 2008). The highest DO value is observed at station 44, and which has the lowest S－0.3 T value (Figure 2). Hard IRM component of station 44 decreases at around 300℃ (Figure 3c). It is suggested that maghemite is aboundant in station 44 sample. As shown in the Day plot and FORC diagrams (Figure 5), magnetic grain-size dependent parameters (Mrs/Ms and Hcr/Hc) indicate that the magnetic grain-size is finer at stations 38 and 43 than at the other stations (Figure 4).The results indicate that oxic bottom water significantly affect to magnetic mineralogy and grain sizes of the sediments. It is known that various substances are transported and cycled in the sediment-water interface (e.g., Berner, 1980). It is inferred that oxygen in overlying bottom water also is transported to the sediments.
TOC contents in the samples from stations 38, 44, 45, and 46 are 2.61-6.43 wt％ (Table 1), and they are high for marine sediments. A large amount of oxygen is consumed for organic matter decompositions, and the sedimentary environment at the station is possibly in an anoxic condition. In general, goethite and maghemite are not stable under such sedimentary environment and dissolve. In spite of the high TOC, goethite and maghemite are present in the studied samples. It is, therefore, possible that differences in the magnetic mineral assemblages in this study area are related to oxic conditions in bottom waters, although the differences may have also been influenced by the supply of sedimentary materials (Table 1).
The highest negative correlation is recognized in the relationship between DO and S－0.3 T (Tables 1 and 2). The composition of magnetic minerals are known to reflect redox condition factors which are TOC and/or DO contents in bottom and/or interstitial water (e.g., Berner 1980; Henshaw and Merill, 1980; Canfield and Berner, 1987; Hilton, 1987; Leslie et al., 1990a, b; Karlin, 1990; Roberts and Turner, 1993; Lui et al., 2004; Garming et al., 2005; Rey et al., 2005; Kawamura et al., 2008, 2012; Rowan et al., 2009; Liu et al., 2012). We suggest that the magnetic mineral assemblage can also be influenced by DO content in bottom waters in this study area. It is implied that the difference in the magnetic mineral assemblages corresponds to the redox condition and reflect the oxidation degree in bottom water in this study area.
In order to describe the magnetic mineral distribution in nearshore sediments, rock magnetic analysis of nearshore sediments was performed at five stations off Chile. DO and DI concentrations in bottom waters were also measured at each station. The intensities of concentration-dependent rock magnetic parameters ($\chi $, $\chi $ARM, SIRM, and HIRM) and magnetic grain size indicators correspond to lithological variations. Results of thermo-magnetometory indicate that magnetite, hematite, (titano)maghemite, and goethite are common in the samples. Maghemite is recognized in the sample which has low S－0.3 T and highest TOC content at most oxic station. Hcr/Hc and Mrs/Ms ratios of the stations plot mostly in the MD region on the Day plot. Higher TOC generally causes anoxic conditions in sediments. However, our results imply that high coercivity magnetic mineral like maghemite is kept under oxic conditions of bottom water in this study area.
We are grateful to Dr. Naomi Harada at JAMSTEC for sample preparations. We thank Dr. Toshiya Kanamatsu, Dr. Takashi Miyazaki, and an anonymous reviewer for valuable comments that helped to improve our manuscript. Funding for this research was partially provided to N.K. by the Nippon foundation. This study was performed under the cooperative research program of Center for Advanced Marine Core Research (CMCR), Kochi University (accept No. 15A007) with the support of JAMSTEC.
