Iron isotopic fractionation during continental weathering

Earth and Planetary Science Letters 228 (2004) 547 – 562 www.elsevier.com/locate/epsl

Iron isotopic fractionation during continental weathering Matthew S. Fantlea,*, Donald J. DePaoloa,b,1 a

University of California, Department of Earth and Planetary Science, Center for Isotope Geochemistry, 307 McCone Hall, Berkeley, CA 94720-4767, United States b MS 90-1160, Earth Sciences Division, E.O. Lawrence Berkeley National Laboratory, Berkeley, CA 94720, United States Received 2 September 2003; received in revised form 15 September 2004; accepted 13 October 2004 Editor: E. Boyle

Abstract The effect of continental weathering on the iron isotope compositions of natural materials is investigated. Unweathered igneous rocks, pelagic clay, and dust fall within the range d 56Fe=0F0.3x. Rivers with large suspended loads also have d 56Fe values near zero. Dilute streams have d 56Fe values that trend towards lower d 56Fe (~
1) suggesting that dissolved riverine iron is isotopically light relative to igneous rocks. Bulk soil and soil leaches display systematically different d 56Fe profiles, indicating that isotopically distinct Fe pools are generated during pedogenesis. Nannofossil ooze, which contains Fe scavenged from the ocean water column, has d 56Fec0, but is consistent with seawater dissolved Fe having negative d 56Fe. It is inferred that continental weathering under modern oxidizing Earth surface conditions preferentially releases dissolved Fe with negative d 56Fe, which is transported in rivers to the ocean. A preliminary analysis of the marine Fe budget suggests that riverine Fe has a substantial role in determining the d 56Fe of both the modern and ancient oceans, but other inputs, particularly that from diagenesis of marine sediments, may also be important. Since the chemical pathways of Fe processing during weathering are dependent on oxidation state and biological activity, Fe isotopes may prove useful for detecting changes in these parameters in the geologic past. D 2004 Elsevier B.V. All rights reserved. Keywords: iron isotopes; weathering; rivers; soils; iron cycle

1. Introduction Recent interest in the iron cycle stems from the role of Fe as a micronutrient in biological systems * Corresponding author. Tel.: +1 510 642 9116; fax: +1 510 642 9520. E-mail addresses: [email protected] (M.S. Fantle)8 [email protected] (D.J. DePaolo). 1 Tel.: +1 510 643 5064; fax: +1 510 642 9520. 0012-821X/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2004.10.013

and its corresponding potential for regulating global climate by fertilizing the ocean and drawing down atmospheric CO2 [1]. To evaluate the role of Fe in climate, the mechanisms by which Fe is transformed and transported in the Earth’s surface environment must be understood. Surface conditions, especially oxidation state, determine the pathways through which Fe is processed during weathering. The modern terrestrial Fe cycle operates in a dominantly oxidizing environment, in which the less soluble

548

M.S. Fantle, D.J. DePaolo / Earth and Planetary Science Letters 228 (2004) 547–562

and generally immobile ferric (III) species is favored [2–6]. Nonetheless, mobilization of Fe under oxidizing conditions can occur by the formation of Fe(III) colloids, solubilization of Fe(III) by organic complexation, and local reduction of Fe(III) to Fe(II). If Fe isotope effects are associated with these reactions, the Fe cycle may be amenable to study using the Fe isotopic composition of natural materials. Differences in the oxidation state of the Earth’s surface environment in the geologic past suggest that Fe was not always processed under oxidizing conditions [7]. These differences may be evident in the Fe isotopic composition of ancient sediments and paleosols [8] and indicative of variations in chemical processing within the Fe cycle. Continental weathering involves the breakdown of rocks and minerals by mechanical and chemical means, the subsequent development of soils, and the transport of weathering products to the ocean by rivers. Recent studies document the isotopic composition of Fe reservoirs and the Fe isotope fractionations produced by certain pathways within the Fe cycle. Igneous rocks have little variation in the 56 Fe/54Fe ratio and, as a result, the Fe isotope composition of igneous rocks (defined as d 56Fe=0) is used as a baseline against which other natural materials are compared [9,10]. Igneous rocks and some clastic sediments have so far been found to have d 56Fe between
0.3x and +0.3x while organic-rich black shales have a d 56Fe range between
2.3x and +0.7x [11,12]. Limited data suggest that exchangeable Fe in soils and sediments, and aqueous Fe in rivers, groundwater, and sediment pore waters can be isotopically fractionated relative to igneous rocks [13,14]. Reduction, complexation with organic ligands, and inorganic speciation of Fe have been shown to affect the isotopic composition of Fe. The reduced or complexed phase can have d 56Fe as much as 1.3x lower than the source Fe [13,15]. Aqueous solutions at equilibrium (22 8C) contain Fe(III) that has d 56Fe values that are 3x higher than that of Fe(II) [16]. Mobile Fe(II) and complexed Fe(III) in weathering environments can be considerably lighter than the material from which they are derived. The extent of isotope fractionation and the degree of mobility are intricately related to the oxidation state,

pH, and both the activity and Fe-affinity of complexing ligands. An additional factor affecting the isotopic composition of Fe in weathering environments is the immobilization of Fe by precipitation and adsorption. Ferric precipitates, such as ferrihydrite and hematite, can have d 56Fe values that are 0.14x higher than aqueous Fe at equilibrium [10,12]. Potentially larger isotope effects may occur in precipitates from a combination of equilibrium and kinetic effects [10,12,14]. Sorption of Fe to charged particle surfaces in soils or natural waters may also sequester Fe with d 56FeN0x, resulting in mobile Fe that has d 56Feb0x [17]. The current work is aimed at evaluating the hypothesis that light Fe is released preferentially during weathering under modern surface Earth conditions. Toward this end, we have measured d 56Fe in river waters, a soil profile, and selected deep-sea sediments to search for evidence that there is systematic Fe isotopic fractionation associated with weathering and related low temperature geochemical processes. A suite of igneous rocks was also measured to verify our choice of a value for d 56Fe=0. Our findings suggest that small amounts of isotopically light Fe are preferentially leached from soil profiles during weathering and transported to the ocean in the dissolved phase (b0.45 Am) of river waters. This result is important when considering the modern isotopic composition of the ocean and differences in d 56Fe of ocean proxies over geologic time.

2. Sample locations and descriptions Igneous rock samples include five volcanic rocks from the Long Valley, California volcanic center [40], the U.S.G.S. rock standard BCR-1 (Columbia River basalt), and a basalt from Mauna Kea volcano cored by the Hawaii Scientific Drilling Project [41,42]. The soil profile chosen for study is from the Ferncreek series, which is developed on a marine terrace near Mendocino, California (Fig. 1). The Ferncreek series (est. age 300 to 400 ka; [18]) consists of deep, somewhat poorly drained soils (clayey, mixed, isomesic Plinthic Haplohumults)

M.S. Fantle, D.J. DePaolo / Earth and Planetary Science Letters 228 (2004) 547–562

developed on lithified marine sediments. The marine sediments are part of the Coastal Belt of the Franciscan complex and are late Cretaceous to early Tertiary in age [19]. Our sample location lies within Bachman’s [19] western unit of the Coastal Belt, a heavily deformed sequence of mudstones, interbedded sandstones and mudstones, and minor conglomerates [19]. The Ferncreek soil contains a ~7 cm organic O-horizon (black in color), a well-developed ~16 cm thick leached E horizon (gray in color), and zones of iron and clay accumulation in the B horizons (~16 to 128 cm depth; orange to red in color). The lower Btv horizons (~67 to 128 cm depth) contain red iron-rich nodules. The dominant vegetation is Bishop pine (Pinus muricata) though Douglas fir (Pseudotsuga menziesii) and redwood (Sequoia sempervivens) also occur in some abundance [20]. The Ferncreek is extremely to strongly acid, with pH between 3.9 (E horizon) and 4.9 [21]. In this region, the mean annual precipitation is ~127 cm and the mean annual air temperature is ~12 8C [21].

!

Fig. 1. Sample location map for eight North American rivers ( ), marine core LL44-GPC3 and DSDP core 590B ( R ), and the Ferncreek soil series (B) whose isotopic compositions are measured in this study.

549

Unfiltered river water samples were collected July 29 through September 24, 2000 from several rivers of the Western United States and Canada (Fig. 1). Most of the rivers drain Mesozoic and Cenozoic volcanic and sedimentary rocks, while two of the sampled rivers (Skagit, Klamath) drain regions that contain significant amounts of Precambrian or Paleozoic rocks. Temperature and pH were measured at each site. Water samples were taken from the uppermost 0.5 m of the river. Stream sediments were collected from the riverbank in clean polyethylene snap-top containers and doublebagged in clean plastic bags. The selected rivers span a range of total Fe concentrations and weathering environments, from the temperate, seasonally wet Northern CaliforniaPacific Northwest climate to the subarctic climate regime of Alaska and Northwest Canada. All are major rivers based on discharge and sediment yield [22]. The western coast of North America contributes an average of 131106 tons year
1 of suspended sediment discharge to the Pacific Ocean while the Yukon contributes a large share of the 444106 tons year
1 of suspended sediment discharge to the Bering Sea [23]. Marine sediment core LL44-GPC3 (30819.9VN, 157849.9VW; 5705 m depth), located in the central gyre of North Pacific Ocean (Fig. 1), is dominated by eolian input [24–28]. The pelagic clay core is 24.31 m in length, the top 20.6 m of which contains a relatively complete Cenozoic record. Sedimentation rates in the core vary between 0.2 and 2 m My
1 while mass accumulation rates are estimated to be between 8 and 137 mg cm
2 ky
1 [25]. Iron content in the core ranges between 49 and 147 mg Fe g
1 sediment, with significant peaks at ~48.3 Ma (147 mg g
1) and 57.3 Ma (129 mg g
1) [25]. Carbonate sediments from Deep Sea Drilling Project Site 590 (Hole B) were sampled on the Lord Howe Rise in the southwest Pacific Ocean (31810.02VS, 163821.51VE; 1299 m water depth) [29,30]. Site 590 is situated on the subtropical divergence in the transitional zone between warm subtropical and temperate regions [29,30]. The sediment is primarily nannofossil ooze but contains foraminifera-rich intervals and thin volcanic ash layers [29,30]. The oldest sediments cored were earliest Miocene (~24 Ma). Mass accumulation rates

550

M.S. Fantle, D.J. DePaolo / Earth and Planetary Science Letters 228 (2004) 547–562

(MAR) at Site 590 were fairly constant throughout the middle and late Miocene at around 2000 g cm
2 My
1 (b63 Am fraction) [31]. Between 6 and 3 Ma, the bulk MAR reached a maximum value ~8000 g cm
2 My
1 (b63 Am fraction) while the carbonate MAR peaked at ~7000 g cm
2 My
1 [31,32]. After 3 Ma, MARs at 590 decreased to the Holocene rate near 1800 g cm
2 My
1.

3. Analytical methods 3.1. Sample preparation Whole rock, bulk soil, unfiltered stream, and bulk sediment samples were dissolved in a heated (b200 8C) solution of concentrated hydrofluoric (HF), perchloric (HClO4), and nitric acids (HNO3). Soil samples were combusted at 600 8C for ~12 to 20 h in a muffle furnace to destroy organic material prior to leaching and bulk dissolution. During ashing, organic complexes are partially broken down and ferrous Fe oxidized to ferric Fe. Selective, nonsequential leaches were performed on the soil samples in water and 0.5 N HCl. Approximately 10 ml of 0.5 N HCl or water was added to 50 mg of pre-combusted soil in acid-washed 30 ml teflon Savillax vials and then shaken at 150 rpm for 24 h prior to filtering under vacuum through acid-washed 0.45 Am Millipore Duraporek polyvinylidene fluoride membrane filters. The goal of the water leach was to mobilize exchangeable Fe and Fe liberated by the breakdown of organic complexes. Because of the low Fe concentration of this fraction, we did not analyze the Fe isotopic composition of the water leach. Since the leaches were nonsequential, the 0.5 N HCl leach contains organic-bound and exchangeable Fe in addition to Fe from soluble amorphous and slightly crystalline oxides and hydroxides. Thus the 0.5 N HCl fraction represents Fe within the soil profile that is most likely to be mobilized (or has already been affected by mobilization processes) during weathering. Marine carbonate sediments were weighed into acid-washed Savillex beakers and dissolved in 0.5 N HCl. The resulting solutions were capped and placed on a hot plate at low heat overnight to ensure total carbonate dissolution. The samples were subse-

quently placed into acid-washed polyethylene centrifuge tubes and centrifuged for 30 min at 4500 rpm. After the supernatant was decanted off, a single 0.5 N HCl rinse was added to the centrifuge tube and the solution recentrifuged and decanted. There was a small amount of detrital noncarbonate material remaining after the second centrifugation step and it is possible that this material may have contributed adsorbed Fe to the bcarbonate fractionQ of the sample. Dilute HCl extraction of Fe from soils and carbonate sediments over a 24-h period was used in place of ammonium oxalate extraction to ensure complete dissolution of amorphous and short-range crystalline Fe [33,34]. While possible that goethite and hematite in the sample were partially dissolved and contributed Fe to the sample [34], goethite and hematite dissolve 10 to 100 times slower than less crystalline oxides such as lepidocrocite and akaganeite in 0.5 N HCl at 25 8C [35]. Hence, the leaching procedure is unlikely to have dissolved more than a small fraction of the goethite and hematite while completely dissolving the less crystalline and amorphous components. Stream sediment fractions were separated from the bulk sediment using the method indicated (Table 1) and then leached in aqua regia prior to Fe separation. The oxidized fraction from the Eel River was composed of rounded sand-size grains with Fe-oxide matrices surrounding lithic/mineral fragments. In all cases, the Fe isotopic composition of the aqua regia leach fraction was measured. 3.2. Chemical separations and mass spectrometry Iron isotope compositions were measured by thermal ionization mass spectrometry (TIMS) with a 54 Fe–58Fe double spike using a VG Sector 54 multicollector mass spectrometer. For each sample, the bspike-subtractedQ 57Fe/56Fe ratio was first determined using an iterative technique to correct for instrumental mass fractionation [36]. The 54Fe/56Fe and 58Fe/56Fe sample ratios were then calculated from the 57Fe/56Fe ratio. Iron in all samples was separated from matrix elements in varying normalities of HNO3 using either a single-column procedure with Eichrom’s RE Spec ion exchange resin or a double-column procedure with Biorad’s AG-50W-X8 cation exchange resin and RE

M.S. Fantle, D.J. DePaolo / Earth and Planetary Science Letters 228 (2004) 547–562 Table 1 (continued)

Table 1 Measured Fe isotopic composition in continental iron cycle Sample measured Mendocino, CA marine terrace soil Bulk Fe O horizon E horizon Btv horizon Bctv horizon 0.5 N HCl leachable Fe O horizon E horizon Btv horizon Bctv horizon

d 56Fe

2r a

(x)

(x)

Hole 590B, 1299 m depth 590B-12-6 590B-47-4

Depth from surface (cm)
0.41
0.61 0.13 0.03

0.16 0.13 0.05 0.04

2 2 5 6


7 to 0 0 to 16 45 to 67 106 to 128+

0.18 0.33
0.52
0.15

0.11 0.09 0.13 0.09

2 2 2 2

Total river water Copper River (Alaska) Susitna River (Alaska) Yukon River (Alaska) Stikine River (British Columbia) Skagit River (Washington) Russian River (California) Klamath River (California) Eel River (California)

[Fe]bulk (ppm)b 21.96


0.49

0.12

2

20.70


0.14

0.06

2

14.01

0.04

0.13

2

3.00


0.14

0.08

2

0.42


0.42

0.09

2

0.169


0.35

0.11

6

0.065


0.64

0.08

4

0.014


0.87

0.10

2

Stream sediments

Separation method

Marine core (30819VN, 157849.9VW)c LL44-GPC3, 5707 m depth 0.25 m depth 2.22 m 6.36 m 12.09 m DSDP core (31810.02VS, 163821.51VE)

d 56Fe

2r a

(x)

(x)

4.14 13.6


0.16
0.08

0.07 0.04

4 3

SiO2 (wt.%) 67.8 67.8 71.9


0.12
0.04 0.00

0.09 0.07 0.08

2 2 2

47.0 50.5 75.1 54.1


0.24
0.05 0.30 0.15

0.08 0.07 0.10 0.08

2 2 2 2

meteorite

0.18 0.04
1.00
0.75
0.89
0.91

0.04 0.10 0.08 0.08 0.09 0.16

2 1 2 2 3 2

Sample measured n


7 to 0 0 to 16 45 to 67 106 to 128+

Eel River Magnetic fraction Oxidized fraction Copper River Magnetic fraction

551

Bulk rocks and minerals Quartz latite (Long Valley) Quartz latite (Long Valley) Deadman rhyolite (Long Valley) Hawaiian basalt Long Valley basalt Long Valley rhyolite BCR-1 (Columbia River basalt) St. Severin chondrite Mojave Desert dustd Ferrihydrite (Montana) Chalcopyrite (Finn)e Sphalerite (Finn)e

Acid mine drainage Deep-sea vent Deep-sea vent

n

a

Franz Hand-pick Hand-magnet

0.05 0.48

0.25 0.27

1 2


0.17

0.07

2


0.11
0.05
0.01
0.13

0.06 0.12 0.04 0.04

3 2 5 6

Age (Ma)c

0.14 0.94 7.6 38.5 Age (Ma)

2r error calculated from the instrumental standard error of the Fe/56Fe ratio. b Concentrations measured by isotope dilution thermal ionization mass spectrometry. c Pelagic clay samples from F. Kyte (UCLA); depths and ages from Kyte et al. [25]. d Mojave (CA) desert soil dust sample courtesy of Dr. Ronald Amundson (UC-Berkeley). e bFinnQ hydrothermal chimney sample courtesy of D. Kelley (UW). 57

Spec resin. While yields from column calibration experiments (using BCR-1) were near 100%, it is possible that the ion exchange chemistry of some natural samples resulted in yields less than 100%. Isotopic fractionation associated with partial yields is accounted for by adding the double spike prior to chemical separation. Processing blanks were less than 0.6% of the total Fe processed per sample. Separated Fe (~4 Ag) was subsequently loaded onto rhenium filaments with SpecPure aluminum, highpurity phosphoric acid, and either colloidal (~0.02 Am particle size) or acid-leached, milled high-purity silica for mass spectrometric analysis. During mass spectrometric analysis, the ion beam intensities of all four Fe nuclides (54, 56, 57 and 58) are measured during one scan. Two additional scans are used to monitor

552

M.S. Fantle, D.J. DePaolo / Earth and Planetary Science Letters 228 (2004) 547–562

either 53Cr+ or 52Cr+ and 60Ni+ for potential isobaric interferences by 54Cr and 58Ni, respectively. The 56Fe beam intensity varies between 1.5 and 2.5 V during the course of a run while the 54Fe+ and 58Fe+(and 57 + Fe ) beams average ~200 and 40 mV, respectively. All Feisotope data are reported in delta notation, ð56=54Þ d56 Fe ¼ ð56=54Þ sample
1 1000, where our standard standard

54

Fe/56Fe ratio is the average value of seven igneous rocks (54Fe/56Fe=0.0636652 or 56Fe/54Fe=15.70717). The 2r error for individual measurements (n=1) is calculated using the instrumental 2r error of the 57 Fe/56Fe ratio, where the 57 beam is the smallest of the four Fe ion beams measured. Our Fe isotope analyses by TIMS have minimum standard errors of 0.0035% (57/56 ratio), which correspond to d 56Fe 2r=0.07x. If the 2r errors for any given analytical run are propagated through the iterative spike subtraction algorithm, the 2r error of d 56Fe increases by a factor of ~2. If the 2r errors are calculated based on a weighted average scheme, the typical 2r error is b0.20x for a sample where the number of replicates, n, is z2. Data from other TIMS studies are reported here relative to BCR-1, BCR-2, or BIR-1, all of which are assumed to be within F0.1x of the igneous rock average [37]. Multiple collector ICP-MS (MC-ICPMS) data are expressed relative to a Columbia River basalt (or equivalently Skulan et al.’s [11] bwholeearth averageQ). We convert Zhu et al.’s [26] e 57FeIRMM-014 data to e 56FeIRMM-014 using the relation e 56FeIRMM-014=0.678e 57FeIRMM-014 [38], and then convert these data to d 56Fewhole-earth using Skulan et al.’s measurement of IRMM-014 relative to their whole-earth average (
0.11F0.07x) [10]. Sharma et al.’s [39] MC-ICPMS Fe isotope data, reported relative to their laboratory standard, are renormalized to BCR-2 using their measurements of IRMM-014 and BCR-2.

4.1. Rocks and minerals Replicate measurements of seven igneous rocks yield a range of 0.5x in d 56Fe (Table 1) [40–42]. The Hawaiian basalt (HSDP core sample R174-9.65) has the lightest Fe isotope composition (
0.24) while the Long Valley rhyolite (LV-7) has the heaviest measured composition (+0.24) among the rocks analyzed. There is no systematic correlation between Fe isotope composition and either silica content or iron content. The average 54Fe/56Fe of the rocks measured is 0.0636652; this ratio is used as a working standard ratio against which all other samples in this study are referenced. Besides igneous rocks, we measured a soil dust sample from the Mojave Desert, the St. Severin chondrite, chalcopyrite (CuFeS2) and sphalerite (ZnS) from the black smoker Finn (East Pacific Rise), and ferrihydrite from an acid mine drainage system in southern Montana. The dust sample (+0.04) does not differ significantly from the rock average, as expected of a material derived from physical weathering of igneous rocks. The St. Severin chondrite has a d 56Fe of +0.18x, slightly heavier than the igneous rock average but within the d 56Fe range of igneous rocks measured in this study. Both the Finn hydrothermal minerals and the acid mine ferrihydrite are isotopically light by
0.75x to
1x. The Finn minerals suggest that the fluids from which they precipitate are isotopically light, in agreement with measurements of hydrothermal fluids (Fig. 2). Sharma et al. [39] report isotopically light hydrothermal fluids in the range
0.7 to
0.2 relative to BCR-2 [39]. It is also possible that there is a kinetic isotope effect associated with the precipitation of hydrothermal minerals within an active deep-sea vent system. This is indicated by studies of d 18O and d 34S in hydrothermal vent fluids, which suggest that the fluids are out of equilibrium with the minerals precipitating from them [43].

4. Results 4.2. Soil profile The results of our Fe isotope measurements in igneous rocks and various minerals, eight North American rivers and some associated stream sediments, a soil profile from the Ferncreek soil series (Mendocino County, California), and marine sediment core samples are shown in Table 1 and Figs. 2–4.

The d 56Fe of iron within the Ferncreek soil profile is between
0.6 and +0.4, with notable differences in d 56Fe between shallow organic-rich and deeper mineral horizons. There is a difference in Fe isotope composition between the bulk shallow horizons and

M.S. Fantle, D.J. DePaolo / Earth and Planetary Science Letters 228 (2004) 547–562

553

Fig. 2. Summary of iron cycle related Fe isotope measurements from this study (5) and recent literature. Delta values are expressed relative to a rock standard, such as BCR-1 (Columbia River basalt) and BIR-1 (Icelandic basalt), or an average value for a measured suite of igneous rocks. Analyses by other authors are indicated by various symbols. The shaded region between ~+0.3x and
0.3x represents the range of d 56Fe measured in igneous rocks and some clastic sediments.

that portion of the shallow horizons leached by 0.5 N HCl. The largest difference in d 56Fe occurs in the leached E horizon (Dleach-bulk=+0.9x). The data suggest that weathering processes operating in these soils can create material with both elevated and decreased d 56Fe relative to the unweathered rock. 4.3. River waters and sediments The d 56Fe values for unfiltered river water samples vary between
0.9 and 0 (Table 1), a range that is skewed towards lighter values relative to igneous rocks

[37]. The three unfiltered river waters with the lowest Fe concentrations, lowest Al/Ca, and Na/Fe and Mg/Fe which approach those of the dissolved (b0.20 Am) load have d 56Fe values of
0.4,
0.6, and
0.9. These same three rivers have relatively high ratios of dissolved Fe to sediment load [44]. In general, the river data confirm that significant Fe isotopic fractionation occurs in weathering processes, producing low values of d 56Fe and suggesting that Fe in the dissolved load is lighter than the Fe of rocks (Fig. 4). Stream bed load fractions from the Eel and Copper Rivers, leached with aqua regia to remove Fe-bearing

554

M.S. Fantle, D.J. DePaolo / Earth and Planetary Science Letters 228 (2004) 547–562

Fig. 3. Chemical characteristics of measured soil profile from the Ferncreek soil series (Mendocino, CA). (a) Loss on ignition (LOI) represents that difference in weight of a given soil sample after combustion at 600 8C for ~12 h and is assumed to signify the amount of organic carbon present in the sample. Iron concentrations (mass Fe per mass soil) in the measured horizons are noted for (b) the bulk soil (%), (c) the 0.5 N HCl leach (ppm), and (d) the double deionized water leach (ppb), as measured by ICP-MS. The Fe isotopic composition (d 56Fe relative to the rock average) of each horizon is shown in (e). Also indicated is a simplified soil profile which denotes the location of the measured horizons, from the uppermost O horizon (black) down to the lowermost Bctv horizon (cross-hatched). Note: the water table fluctuates seasonally so that the Bctv horizon is saturated in the winter and dry in the summer.

M.S. Fantle, D.J. DePaolo / Earth and Planetary Science Letters 228 (2004) 547–562

Fig. 4. Isotopic composition of riverine Fe (d 56Fe, x) as a function of Fe concentration in the bulk waters.

coatings on the silicate grains and dissolve oxides, yield leachate d 56Fe of
0.17 to +0.48. The two magnetic fractions yield d 56Fe similar to igneous rocks while the oxidized sediment fraction from the Eel River is ~0.5x heavier than the rock average. These data suggest that primary minerals which are essentially unaffected by weathering do not display fractionations from igneous rocks while secondary Feoxides, formed during the creation of soils or while transported in fluvial systems, may be isotopically heavy. 4.4. Marine sediments Four samples of deep-sea pelagic clay, which range in age from 0.14 to 38.5 Ma, have uniform d 56Fe of
0.13x to
0.01x. The clay d 56Fe values fall within the range of values for Chinese Xifeng loess (+0.24 and
0.13), Xifeng paleosols (0.00 and +0.10x), Mojave Desert dust (+0.04x), and igneous rocks (Table 1; [45]), indicating that these finegrained clastic materials contain Fe that is unfractionated with respect to igneous rocks, as expected if they are derived from igneous rocks primarily by physical weathering [37]. Assumption of an Asian dust source is supported by analyses of the present-day dust cycle, in which

555

dust from Asia regularly reaches the central Pacific in fairly high concentrations [46]. Total inversion end member modeling by Kyte et al. [25] indicates that N80% of marine core LL44-GPC3 is composed of eolian material. In addition, Hf and Nd isotope studies of core LL44-GPC3 and neighboring sediments (ODP sites 885/886 at 44.78N, 168.38W) indicate that basins north of the Tibetan Plateau and the Gobi Desert, and not North or South America, are the most probable source material for the dust found in core LL44-GPC3 [47,48]. Mass fluxes to core LL44-GPC3 are estimated to be 10–20 mg cm
2 ky
1 until the end of the Paleogene (~24 Ma). About 25 Ma, the mass flux doubled and increased further at ~3 Ma to 30–250 mg cm
2 ky
1 [49]. At this time, additional eolian material may also have been provided to core LL44-GPC3 from the Kamchatka peninsula [47]. Given these assumptions, the observation of coincident d 56Fe in pelagic clays, Asian loess, and igneous rocks suggests that most of the source Fe is unfractionated by alteration processes during continental weathering, transport, and deposition. This hypothesis is supported by the mineralogy of nearby sediments from ODP sites 885/886, which suggest that the degree of chemical weathering in the eolian component of the sediments is low over the past 11 Ma [47]. The d 56Fe values of carbonate ooze (
0.08 and
0.16) from the southwest Pacific are similar to the pelagic clays and also within the range of igneous rocks measured in this study (Fig. 1, Table 1). The majority of the iron measured in the carbonate oozes derives from Fe-oxides precipitated on the exterior of the carbonate tests in the water column and postdepositional Fe-Mn coatings. If, as suggested by experiments, Fe hydroxides tend to be higher in d 56Fe than the dissolved Fe pool from which they form, then the d 56Fe of marine carbonates suggest that the marine dissolved Fe in the West Pacific has negative d 56Fe [10,12,14].

5. Discussion The observations that we consider to be significant for analyzing the behavior of Fe isotopes in the continental weathering cycle are: (1) intraprofile

556

M.S. Fantle, D.J. DePaolo / Earth and Planetary Science Letters 228 (2004) 547–562

fractionation in the Ferncreek soil, (2) the range of d 56Fe in river waters, including negative values, and (3) the low d 56Fe of the Eel River water sample that we consider most representative of dissolved riverine Fe. Considering these observations in the context of other recently reported experimental and field studies (Fig. 2), we hypothesize that Fe isotope fractionation during weathering is produced by cycles of Fe reduction, organic-Fe chelation, biological cycling, and hydroxide precipitation, each of which has been documented to affect Fe isotope composition to some degree [10,13–15,38]. 5.1. Fractionation of Fe isotopes by soil processes Recent experimental work suggests that isotopically light pools of Fe may be generated in soils via reductive dissolution of Fe(III) solids, solubilization of Fe(III) from solids via organic complexation, and/ or preferential sorption of isotopically heavy Fe to particle surfaces [13,15,17]. Reservoirs of relatively heavy Fe may then be created through oxidation of Fe(II), equilibrium fractionation between Fe(II) and Fe(III) in solution, and precipitation of Fe(III) as oxides and/or hydroxides. The experimental work has been supported by limited measurements in the field, where soil exchangeable Fe (d 56Fe=
0.85x) and soil Fe-oxides (d 56Fe=+0.16x) display the expected behavior [13]. The complexity and general nature of processes within a soil profile are illustrated by the data from the Ferncreek soil (Fig. 3). The bulk O and E horizons have low Fe concentrations and relatively light d 56Fe in comparison to the deeper B horizons which have higher Fe concentrations and d 56Fe values slightly greater than zero. The data suggest that organic Fe is significantly lighter than the mineral Fe in soils and that either plant uptake or mineral weathering processes preferentially release light Fe that can then be taken up by the biosphere. It is also likely that the lower mineral horizons are affected by a combination of processes, such as precipitation of secondary Feoxides, reductive dissolution, and biological uptake, each of which acts to increase the d 56Fe of these horizons. Iron isotope fractionation during biological uptake of Fe can be understood in the context of organic complexation experiments and the mechanisms by

which plants obtain Fe. The use of Fe by plant root systems requires Fe to be organically chelated prior to uptake [50,51]. In the case of grasses, roots release bphytosiderophoresQ into the soil and siderophorebound Fe is subsequently taken up by plant roots [51]. If light Fe is preferentially complexed by phytosiderophores, as suggested by Brantley et al. [13], then plant tissue will be isotopically lighter than the bulk soil from which it was derived. Hence, the O horizon, composed primarily of decomposed organic matter derived from surface plant material, has a bulk d 56Fe that is
0.5x relative to the Fe in the B horizons (Table 1; Fig. 3e). Most isotopically light Fe in the O horizon is organically bound, most likely to large organic molecules derived from the decay of surficial plant debris. Equilibrium models of Fe speciation in soils indicate that there is essentially no free Fe ion in soil solutions; instead Fe is bound primarily to high molecular weight organic acids with a small proportion complexed by oxalate and citrate [52]. The Fepoor E horizon also has low d 56Fe, suggesting that a large proportion of Fe in the E horizon is derived from the mobile, organically bound Fe transported from the O horizon. The 0.5 N HCl soil leachates (Fig. 3) represent adsorbed Fe, organic Fe, and oxide- or hydroxidebound Fe with minimal silicate-bound and crystalline oxide Fe [33]. The Fe from the HCl leach is slightly heavy in the O and E horizons and lighter in the B horizons. Mass balance calculations indicate that the O and E horizons contain bunleachedQ components that are identical (
0.60x and
0.66x, respectively). This suggests that there is a component in the soil that is isotopically light, makes up ~75% to 95% of the O and E horizons, and is resistant to dissolution (size fraction N0.45 Am) even after ashing. This fraction may be organic, especially if ashing creates insoluble residues greater than 0.45 Am. Other authors have noted that soils subjected to elevated temperatures (90–560 8C) may increase their particle size via flocculation, aggregation, and cementation [53,54]. In addition, it is possible for soil organic matter to survive at high temperatures (560 8C) [53]. The initial Fe sources to the soil are unweathered silicates and oxides, a bprimaryQ reservoir that we assume has a starting d 56Fe ~0x. Some ligands preferentially complex isotopically light Fe. If these complexes are subsequently taken up by plants, then

M.S. Fantle, D.J. DePaolo / Earth and Planetary Science Letters 228 (2004) 547–562

this surface vegetation can have d 56Fe values less than zero. If bprimaryQ Fe, supplied by weathering minerals, is continuously transferred to the mobile reservoir, then the d 56Fe of the mobile reservoir will not change. However, if the supply of Fe to the mobile reservoir occurs in pulses, corresponding to seasonal wet–dry cycles for example, the mobile reservoir may be sufficiently drawn down so that the d 56Fe of the mobile reservoir will change. In the bclosed systemQ scenario, the mobile reservoir will become heavy as light Fe is removed and sequestered in vegetation. This trend is similar to the one we see in Fig. 3, where the E horizon has the smaller and heavier mobile Fe reservoir relative to the O horizon. Additional explanations for the heavy character of the leach are fractionation during either precipitation or equilibrium speciation between aqueous Fe(III) and Fe(II) prior to precipitation. The isotopic composition of the lower mineral horizons is affected by reductive dissolution of Fehydroxide coatings under saturated conditions [55]. During reductive dissolution, isotopically light Fe may be released into solution, creating a mobile reservoir which is isotopically lighter than the parent material and a bulk profile that is slightly heavier than the assumed parent material. This is exactly the pattern observed in the Btv horizon, in which the 0.5 N HCl leach is about 0.5x lighter and the bulk soil about 0.2x heavier than the assumed 0x parent material (Fig. 3e).

557

complexed and exchangeable Fe, tends to have relatively low d 56Fe (Fig. 3; [13]). In addition, experimental data suggest that, at equilibrium, mobile Fe(II) is lighter by
3x relative to immobile Fe(III) [16,56]. Adsorbed Fe may also be significantly heavier than non-adsorbed Fe by +0.8x, leaving a mobile phase lighter than the residual material [17]. These data suggest that the dissolved Fe supplied to rivers from rock weathering could have low d 56Fe, as a result of isotope fractionation during equilibrium speciation and sorption to particles in soils. As relatively acidic soil water (pHb5) mixes with river water (pH 6–7), further isotope fractionation may occur due to changing redox speciation and increased sorption onto colloidal particles. Precipitation of Fe in streams may also sequester heavy Fe(III) in secondary oxides, driving the d 56Fe of the dissolved Fe to lower values [14]. Kinetic calculations indicate that Fe(III) is favored over Fe(II) in oxidized, circumneutral environments due to rapid oxidation of Fe(II) [57–59]. Yet in natural systems exposed to sunlight, Fe(II) oxidation by O2 and H2O2 can be balanced by photochemical Fe(III) reduction in sunlit surface waters, creating a steady state Fe(II) reservoir under oxidizing conditions [59– 61]. Rivers are particularly favorable environments for relatively large steady-state Fe(II) reservoirs, due to significantly slower rates of Fe(II) oxidation compared to higher ionic strength solutions such as the surface ocean [57,62,63]. 5.3. The global Fe budget

5.2. Isotopic composition of riverine Fe The water samples measured in this study represent the total Fe carried by rivers in the suspended and dissolved fractions. Rivers with high particle loads, such as the Yukon and Susitna, have d 56Fe similar to our rock average. This is in agreement with previous measurements of riverine suspended loads, which suggest d 56Fe of suspended sediment between
0.3x and 0.3x [12]. Rivers with lower particle loads, such as the Klamath and Eel, have d 56Fe decidedly lower than our rock average (Fig. 4). If we assume that the suspended load has d 56Fec0F0.3x, this result suggests that dissolved Fe is isotopically light. The inference that riverine dissolved Fe can have d 56Feb0x is consistent with other observations. Soil data indicate that soluble Fe, such as organically

The presence of isotopically light dissolved Fe in rivers has implications for terrestrial input to the ocean and the global Fe cycle. To understand the implications, we assemble a preliminary global Fe budget from available annual Fe flux data (Fig. 5). Rivers transport 12 to 18 Tmol Fe year
1, but only about 0.026 Tmol year
1 is in the dissolved form (operationally defined as b0.45 Am) [64]. The dissolved fraction therefore includes colloids as well as truly dissolved Fe. As much as 90% of the riverine dissolved Fe is removed in estuaries by aggregation and sinking of colloids, thus the minimum net riverine Fe flux to the oceans is about 0.003 Tmol year
1 [64,65]. Marine hydrothermal systems [64,66] and the atmosphere [64,67] contribute less total Fe to the global ocean than rivers, but the dissolved Fe fluxes

558

M.S. Fantle, D.J. DePaolo / Earth and Planetary Science Letters 228 (2004) 547–562

Fig. 5. Two-box model of the ocean system, indicating total and dissolved Fe fluxes for the three primary inputs to the ocean and estimated sedimentary and reduction Fe fluxes. At steady state, given a total Fe input to the global ocean between 12.8 and 18.7 Tmol Fe year
1, we estimate an average global Fe sedimentation flux of 40 to 50 mmol Fe m
2 year
1.

from these two sources may be similar in magnitude to the dissolved riverine flux [68,69]. The flux of diagenetic Fe from marine sediments to the ocean may also be significant [70–72]. Iron is reduced during burial in marine sediments [73], and the escape of Fe from sediments into the overlying water column correlates with, and is presumably controlled by, the flux of organic carbon (Corg) to the sediments [72,74]. Diagenetic Fe fluxes of 12 to 35 mmol Fe m
2 year
1 are estimated for sediments receiving z1200 mmol C m
2 year
1 [72,74] while direct flux measurements of ~18 mmol Fe m
2 year
1 are observed in Black Sea shelf sediments [75]. Diagenetic Fe reduction is likely to be restricted to coastal regions and enclosed sedimentary basins which can have Corg accumulation rates between 500

and 4000 mmol C m
2 year
1 [71,76]. Given a coastal ocean surface area of 31013 m2 (0 to 200 m depth) [77], these numbers suggest that the diagenetic Fe flux could be as high as 0.35 to 1 Tmol Fe year
1, much larger than other sources of marine dissolved Fe. The influence of diagenetic Fe on the global Fe cycle depends on the proportion of diagenetic Fe that survives in the water column. Ferrous Fe oxidized by O2 in surface seawater (T~20–25 8C) has a short calculated half-life of ~1–19 min (pH 7.6–8.3); in bottom waters (T~5 8C), Fe(II) has a longer half-life of ~4–10 h (pH 7.5–7.7) [6,57]. Experimental work has shown that some organic ligands may decrease the oxidation rate of Fe(II) in seawater [78,79]. Therefore, some portion of the diagenetically released Fe(II) may be stabilized by organic complexation and low temperatures, effectively increasing the residence time of the Fe(II) in the water column. Since the ultimate fate of Fe(II) is oxidation to Fe(III) in bottom waters devoid of photochemical reduction, solubility arguments constrain the dissolved flux of diagenetic Fe. The solubility of Fe(III) in the modern ocean is generally thought to be between 0.1 and 1.5 nmol l
1 but has been found to be as high as 5–10 nmol l
1 in the North Pacific Ocean, Sargasso Sea, and coastal Pacific and regions of high productivity [6,80–84]. Our total estimated diagenetic flux exceeds the range of Fe(III) solubilities in the coastal ocean. Based on these studies of Fe(III) solubility, only a small amount of diagenetically released Fe (0.1–17%) survives removal from the water column. Applied to our total estimated diagenetic Fe flux (0.35 to 1 Tmol Fe year
1), we arrive at a dissolved flux (b0.17 Tmol Fe year
1) similar in magnitude to the other dissolved Fe fluxes (Fig. 5). The flux of diagenetic Fe to the open ocean from coastal areas is uncertain, though it is likely that this flux is episodic and a function of local productivity. However, our preliminary analysis suggests that the flux of diagenetic Fe could be significant (if not large) in comparison to the other fluxes. The diagenetic Fe flux may prove particularly important during regional marine or global anoxic events, when decreased rates of Fe(II) oxidation enhance the flux of diagenetic Fe that reaches the deep ocean. Dissolved Fe concentrations in the modern ocean vary both spatially and temporally [85–87]. The residence time of dissolved marine Fe calculated using

M.S. Fantle, D.J. DePaolo / Earth and Planetary Science Letters 228 (2004) 547–562

the dissolved fluxes in Fig. 5 and estimates of the dissolved ocean Fe reservoir (100%, 50%, and 20% of the total Fe reservoir) are in the range ~1 to 370 years, far less than the mixing time of the ocean (~1000 years) [88]. Thus we expect the modern ocean to be isotopically heterogeneous. An isotopically wellmixed ocean may only occur when the input or output fluxes are substantially smaller or solubility constraints change to allow for a larger dissolved Fe reservoir.

6. Conclusions We suggest that the net effect of continental weathering processes is to mobilize small amounts of isotopically light Fe in an bexchangeableQ, or dissolved, form. Rivers transport this small isotopically light Fe reservoir in solution to the ocean, leaving behind continental residues which are isotopically similar to or slightly heavier than the parent material. Under the oxidizing conditions that characterize the modern Earth surface environment, dissolved Fe in the oceans should be relatively light, but variable, reflecting the influence of dissolved riverine Fe and a relatively short residence time. A major uncertainty in the modern global Fe cycle is the magnitude and isotopic composition of the flux of diagenetic Fe from sediments to the oceans. Processes within soils produce both isotopically light and heavy Fe. Biological processes, such as the growth of surface vegetation and synthesis of organic ligands, produce a source of isotopically light, relatively mobile Fe within the soil that can be transported to streams via soil pore waters. Local reductive dissolution of oxide coatings in watersaturated horizons, while subject to seasonal and annual fluctuations, can also contribute isotopically light Fe to the mobile phase. In rivers, additional isotope fractionation of Fe may take place by equilibrium speciation, precipitation of colloidal Fe(III), and sorption of dissolved Fe onto particle surfaces. Upon entering the ocean, particulate Fe (including colloidal material) aggregates and sediments, resulting in a significant and isotopically light Fe input to the ocean [65]. In the modern oceans, the residence time of dissolved Fe is short and hence the isotopic composition is likely to be spatially variable. In the oceans of

559

the geologic past, different oxidation state and weathering rates could have produced more isotopically homogeneous oceans. Fe isotopes may be useful as a paleo-pO2 and biological activity proxy in both paleosols and marine deposits, but the spatial variability of oxidation potential and biological activity will make it difficult to draw conclusions about global conditions.

Acknowledgements This work was partially supported by a NASA ESS Graduate Fellowship to MSF and by NSF Grant EAR9909639. Major funding was provided by the Director, Office of Science, Basic Energy Sciences, Chemical Sciences, Biosciences, and Geosciences Division of the U.S. Department of Energy under Contract No. De-AC03-76SF00098. Samples of core LL44-GPC3 were provided by Dr. Frank Kyte (UCLA). Tom Owens helped with mass spectrometric analyses. The authors also thank Ryan Petterson for assistance with river sampling and Ron Amundson for advice and help with soil sampling.

References [1] J.H. Martin, Glacial–interglacial CO 2 change: the iron hypothesis, Paleoceanography 5 (1) (1990) 1 – 13. [2] X.W. Liu, F.J. Millero, The solubility of iron in seawater, Mar. Chem. 77 (2002) 43 – 54. [3] X.W. Liu, F.J. Millero, The solubility of iron hydroxide in sodium chloride solutions, Geochim. Cosmochim. Acta 63 (1999) 3487 – 3497. [4] F.J. Millero, Solubility of Fe(III) in seawater, Earth Planet. Sci. Lett. 154 (1–4) (1998) 323 – 329. [5] F.J. Millero, W.S. Yao, J. Aicher, The speciation of Fe(II) and Fe(III) in natural waters, Mar. Chem. 50 (1–4) (1995) 21 – 39. [6] K. Kuma, J. Nishioka, K. Matsunaga, Controls on iron(III) hydroxide solubility in seawater: the influence of pH and natural organic chelators, Limnol. Oceanogr. 41 (3) (1996) 396 – 407. [7] J.F. Kasting, Earth’s early atmosphere, Science 259 (5097) (1993) 920 – 926. [8] G.L. Arnold, S. Weyer, A.D. Anbar, Fe isotope variations in natural materials measured using high mass resolution multiple collector ICPMS, Analyt. Chem. 76 (2004) 322 – 327. [9] B.L. Beard, C.M. Johnson, High precision iron isotope measurements of terrestrial and lunar materials, Geochim. Cosmochim. Acta 63 (11–12) (1999) 1653 – 1660.

560

M.S. Fantle, D.J. DePaolo / Earth and Planetary Science Letters 228 (2004) 547–562

[10] J.L. Skulan, B.L. Beard, C.M. Johnson, Kinetic and equilibrium Fe isotope fractionation between aqueous Fe(III) and hematite, Geochim. Cosmochim. Acta 66 (17) (2002) 2995 – 3015. [11] C.M. Johnson, B.L. Beard, N.J. Beukes, C. Klein, J.M. O’Leary, Ancient geochemical cycling in the Earth as inferred from Fe isotope studies of banded iron formations from the Transvaal Craton, Contrib. Mineral. Petrol. 144 (2003) 523 – 547. [12] B.L. Beard, C.M. Johnson, Fe isotope variations in the modern and ancient Earth and other planetary bodies, in: Geochemistry of non-traditional stable isotopes, Rev. Mineral. Geochem. 55 (2004) 319 – 357. [13] S.L. Brantley, L. Liermann, T.D. Bullen, Fractionation of Fe isotopes by soil microbes and organic acids, Geology 29 (6) (2001) 535 – 538. [14] T.D. Bullen, A.F. White, C.W. Childs, D.V. Vivit, M.S. Schulz, Demonstration of significant abiotic iron isotope fractionation in nature, Geology 29 (8) (2001) 699 – 702. [15] B.L. Beard, C.M. Johnson, L. Cox, H. Sun, K.H. Nealson, C. Aguilar, Iron isotope biosignatures, Science 285 (5435) (1999) 1889 – 1892. [16] S.A. Welch, B.L. Beard, C.M. Johnson, P.S. Braterman, Kinetic and equilibrium Fe isotope fractionation between aqueous Fe(II) and Fe(III), Geochim. Cosmochim. Acta 67 (22) (2003) 4231 – 4250. [17] G.A. Icopini, A.D. Anbar, S.S. Ruebush, M. Tien, S.L. Brantley, Iron isotope fractionation during microbial reduction of iron: the importance of adsorption, Geology 32 (3) (2004) 205 – 208. [18] R. Amundson, in: M.S. Fantle (Ed.), Relative Marine Terrace Ages of Mendocino Soil Chronosequence, Berkeley, 2003. [19] S.B. Bachman, Sedimentation and margin tectonics of the Coastal Belt Franciscan, Mendocino Coast, northern California, Doctoral dissertation, University of California, 1979. [20] R.E. Sholars, The pygmy forest and associated plant communities of coastal Mendocino County, California: genesis– vegetation–soil, ix, 50 pp., R.E. Sholars, Mendocino, Calif. (44650 Surfwood Dr., Mendocino 95460), 1982. [21] J. Rittiman, A. Carl, T. Thorson, Soil Survey of Mendocino County, California, Western Part 2003, Natural Resources Conservation Service, 1998. [22] J.D. Milliman, J.P.M. Syvitski, Geomorphic tectonic control of sediment discharge to the ocean—the importance of small mountainous rivers, J. Geol. 100 (5) (1992) 525 – 544. [23] J.D. Milliman, R.H. Meade, World-wide delivery of river sediment to the oceans, J. Geol. 91 (1) (1983) 1 – 21. [24] B.H. Corliss, C.D. Hollister, A paleoenvironmental model for Cenozoic sedimentation in the central North Pacific, in: R.A. Scrutton, M. Talwani (Eds.), The Ocean Floor, Wiley & Sons, 1982, pp. 277 – 304. [25] F.T. Kyte, M. Leinen, G.R. Heath, L. Zhou, Cenozoic sedimentation history of the central North Pacific—inferences from the elemental geochemistry of core Ll44-Gpc3, Geochim. Cosmochim. Acta 57 (8) (1993) 1719 – 1740. [26] M. Leinen, G.R. Heath, Sedimentary indicators of atmospheric activity in the Northern Hemisphere during the

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

Cenozoic, Palaeogeogr. Palaeoclimatol. Palaeoecol. 36 (1–2) (1981) 1 – 21. D.K. Rea, T.R. Janecek, Late Cenozoic changes in atmospheric circulation deduced from north pacific eolian sediments, Mar. Geol. 49 (1–2) (1982) 149 – 167. T.R. Janecek, D.K. Rea, Eolian deposition in the northeast Pacific-Ocean—Cenozoic history of atmospheric circulation, Geol. Soc. Am. Bull. 94 (6) (1983) 730 – 738. J.P. Kennett, C.C. von der Borch, P.A. Baker, C.E. Barton, A. Boersma, J.P. Caulet, W.C. Dudley, Jr., J.V. Gardner, D.G. Jenkins, W.H. Lohman, E. Martini, R.B. Merrill, R.H. Morin, C.S. Nelson, C. Robert, M.S. Srinivasan, R. Stein, A. Takeuchi, Introduction and objectives, techniques, and explanatory notes, in: Initial reports of the Deep Sea Drilling Project covering Leg 90 of the cruises of the drilling vessel Glomar Challenger; Noumea, New Caledonia, to Wellington, New Zealand, December 1982–January 1983; Part 1, J.H. Blakeslee, ed. 90, pp. 3–16, U.S. Govt. Printing Office, Washington, 1986. J.P. Kennett, C.C. von der Borch, P.A. Baker, C.E. Barton, A. Boersma, J.P. Caulet, W.C. Dudley, Jr., J.V. Gardner, D.G. Jenkins, W.H. Lohman, E. Martini, R.B. Merrill, R.H. Morin, C.S. Nelson, C. Robert, M.S. Srinivasan, R. Stein, A. Takeuchi, Site 590: Lord Howe Rise, 31 8S, in: Initial reports of the Deep Sea Drilling Project covering Leg 90 of the cruises of the drilling vessel Glomar Challenger; Noumea, New Caledonia, to Wellington, New Zealand, December 1982– January 1983; Part 1, J.H. Blakeslee, ed. 90, pp. 263–376, U.S. Govt. Printing Office, Washington, 1986. J.P. Kennett, C.C. von der Borch, Southwest Pacific Cenozoic paleoceanography, in: Initial reports of the Deep Sea Drilling Project covering Leg 90 of the cruises of the drilling vessel Glomar Challenger; Noumea, New Caledonia, to Wellington, New Zealand, December 1982–January 1983; Part 1, J.H. Blakeslee, ed. 90, pp. 1493–1517, U.S. Govt. Printing Office, Washington, 1986. K.M. Grant, G.R. Dickens, Coupled productivity and carbon isotope records in the southwest Pacific Ocean during the late Miocene–early Pliocene biogenic bloom, Palaeogeogr. Palaeoclimatol. Palaeoecol. 187 (1–2) (2002) 61 – 82. R.A. Sutherland, Comparison between non-residual Al, Co, Cu, Fe, Mn, Ni, Pb and Zn released by a three-step sequential extraction procedure and a dilute hydrochloric acid leach for soil and road deposited sediment, Appl. Geochem. 17 (4) (2002) 353 – 365. R.J. Heck, A.R. Mermut, Mechanisms and rates of iron dissolution from Brazilian Plinthustalfs by dilute hydrochloric and ascorbic acids, Soil Sci. 164 (4) (1999) 242 – 251. P.S. Sidhu, R.J. Gilkes, R.M. Cornell, A.M. Posner, J.P. Quirk, Dissolution of iron-oxides and oxyhydroxides in hydrochloric and perchloric acids, Clays Clay Miner. 29 (4) (1981) 269 – 276. T.M. Johnson, M.J. Herbel, T.D. Bullen, P.T. Zawislanski, Selenium isotope ratios as indicators of selenium sources and oxyanion reduction, Geochim. Cosmochim. Acta 63 (18) (1999) 2775 – 2783.

M.S. Fantle, D.J. DePaolo / Earth and Planetary Science Letters 228 (2004) 547–562 [37] B.L. Beard, C.M. Johnson, J.L. Skulan, K.H. Nealson, L. Cox, H. Sun, Application of Fe isotopes to tracing the geochemical and biological cycling of Fe, Chem. Geol. 195 (2003) 87 – 117. [38] A. Matthews, X.-K. Zhu, K. O’Nions, Kinetic iron stable isotope fractionation between iron (-II) and (-III) complexes in solution, Earth Planet. Sci. Lett. 192 (2001) 81 – 92. [39] M. Sharma, M. Polizzotto, A.D. Anbar, Iron isotopes in hot springs along the Juan de Fuca ridge, Earth Planet. Sci. Lett. 6036 (2001) 1 – 13. [40] R.A. Bailey, G.B. Dalrymple, M.A. Lanphere, Volcanism, structure, and geochronology of Long-Valley Caldera, MonoCounty, California, J. Geophys. Res. 81 (5) (1976) 725 – 744. [41] E.M. Stolper, D.J. DePaolo, D.M. Thomas, Introduction to special section: Hawaii scientific drilling project, J. Geophys. Res.-Solid Earth 101 (B5) (1996) 11593 – 11598. [42] D.J. Depaolo, E.M. Stolper, D.M. Thomas, Deep drilling into a Hawaiian volcano, EOS Trans. AGU 82 (149) (2001) 154 – 155. [43] H. Chiba, N. Uchiyama, D.A.H. Teagle, Stable isotope study of anhydrite and sulfide minerals at the TAG hydrothermal mound, Mid-Atlantic Ridge, 268N, Proc. ODP, Sci. Results 158 (1998) 85 – 90. [44] U.S.G.S., Water-Quality Data for USA 2002. [45] X. Zhu, K. O’Nions, Y. Guo, B.C. Reynolds, Secular variation of iron isotopes in North Atlantic Deep Water, Science 287 (2000) 2000 – 2002. [46] J.M. Prospero, P. Ginoux, O. Torres, S.E. Nicholson, T.E. Gill, Environmental characterization of global sources of atmospheric soil dust identified with the Nimbus 7 Total Ozone Mapping Spectrometer (TOMS) absorbing aerosol product, Rev. Geophys. 40 (1) (2002) 1 – 31. [47] T. Pettke, A.N. Halliday, C.M. Hall, D.K. Rea, Dust production and deposition in Asia and the North Pacific Ocean over the past 12 Myr, Earth Planet. Sci. Lett. 178 (3–4) (2000) 397 – 413. [48] T. Pettke, D.C. Lee, A.N. Halliday, D.K. Rea, Radiogenic Hf isotopic compositions of continental eolian dust from Asia, its variability and its implications for seawater Hf, Earth Planet. Sci. Lett. 202 (2) (2002) 453 – 464. [49] D.K. Rea, The paleoclimatic record provided by eolian deposition in the deep-sea—the geologic history of wind, Rev. Geophys. 32 (2) (1994) 159 – 195. [50] H. Siebner-Freibach, Y. Hadar, Y. Chen, Siderophores sorbed on Ca-montmorillonite as an iron source for plants, Plant Soil 251 (1) (2003) 115 – 124. [51] R. Hell, U.W. Stephan, Iron uptake, trafficking and homeostasis in plants, Planta 216 (4) (2003) 541 – 551. [52] P.A.W. van Hees, U.S. Lundstrom, Equilibrium models of aluminium and iron complexation with different organic acids in soil solution, Geoderma 94 (2–4) (2000) 201 – 221. [53] G. Giovannini, S. Lucchesi, Modifications induced in soil physico-chemical parameters by experimental fires at different intensities, Soil Sci. 162 (7) (1997) 479 – 486. [54] D. Badia, C. Marti, Plant ash and heat intensity effects on chemical and physical properties of two contrasting soils, Arid Land Res. Manage. 17 (1) (2003) 23 – 41.

561

[55] F. Favre, D. Tessier, M. Abdelmoula, J.M. Genin, W.P. Gates, P. Boivin, Iron reduction and changes in cation exchange capacity in intermittently waterlogged soil, Eur. J. Soil Sci. 53 (2) (2002) 175 – 183. [56] C.M. Johnson, J.L. Skulan, B.L. Beard, H. Sun, K.H. Nealson, P.S. Braterman, Isotopic fractionation between Fe(III) and Fe(II) in aqueous solutions, Earth Planet. Sci. Lett. 195 (1–2) (2002) 141 – 153. [57] F.J. Millero, S. Sotolongo, M. Izaguirre, The oxidationkinetics of Fe(II) in seawater, Geochim. Cosmochim. Acta 51 (4) (1987) 793 – 801. [58] F.J. Millero, S. Sotolongo, The oxidation of Fe(II) with H2O2 in seawater, Geochim. Cosmochim. Acta 53 (8) (1989) 1867 – 1873. [59] D.W. King, H.A. Lounsbury, F.J. Millero, Rates and mechanism of Fe(II) oxidation at nanomolar total iron concentrations, Environ. Sci. Technol. 29 (1995) 818 – 824. [60] F.J. Millero, in: The Physical Chemistry of Natural Waters, vol. xix, Wiley-Interscience, New York, 2001 (654 pp.). [61] B.M. Voelker, D.L. Sedlak, Iron reduction by photoproduced superoxide in seawater, Mar. Chem. 50 (1–4) (1995) 93 – 102. [62] D.W. King, R.A. Aldrich, S.E. Charnecki, Photochemical redox cycling of iron in NaCl solutions, Mar. Chem. 44 (2–4) (1993) 105 – 120. [63] F.J. Millero, Chemical Oceanography, CRC Press, Boca Raton, 1996, 469 pp. [64] S.W. Poulton, R. Raiswell, The low-temperature geochemical cycle of iron: from continental fluxes to marine sediment deposition, Am. J. Sci. 302 (9) (2002) 774 – 805. [65] H.J.W. de Baar, J.T.M. de Jong, Distributions, sources and sinks of iron in seawater, in: D.R. Turner, K.A. Hunter (Eds.), The Biogeochemistry of Iron in Seawater, IUPAC Series on Analytical and Physical Chemistry of Environmental Systems, vol. 7, J. Wiley, Chichester, NY, 2001, pp. 123 – 253. [66] H. Elderfield, A. Schultz, Mid-ocean ridge hydrothermal fluxes and the chemical composition of the ocean, Annu. Rev. Earth Planet. Sci. 24 (1996) 191 – 224. [67] T.D. Jickles, L.J. Spokes, Atmospheric iron inputs to the oceans, in: D.R. Turner, K.A. Hunter (Eds.), The Biogeochemistry of Iron in Seawater, IUPAC Series on Analytical and Physical Chemistry of Environmental Systems, vol. 7, J. Wiley, Chichester, NY, 2001, pp. 85 – 121. [68] I.Y. Fung, S.K. Meyn, I. Tegen, S.C. Doney, J.G. John, J.K.B. Bishop, Iron supply and demand in the upper ocean, Glob. Biogeochem. Cycles 14 (1) (2000) 281 – 295. [69] J.K. Moore, S.C. Doney, D.M. Glover, I.Y. Fung, Iron cycling and nutrient-limitation patterns in surface waters of the World Ocean, Deep-Sea Research. Part II. Top. Stud. Oceanogr. 49 (1–3) (2002) 463 – 507. [70] D.J. Mackey, J.E. O’Sullivan, R.J. Watson, Iron in the western Pacific: a riverine or hydrothermal source for iron in the equatorial undercurrent? Deep-Sea Research. Part I. Oceanograph. Res. Pap. 49 (5) (2002) 877 – 893. [71] K.S. Johnson, F.P. Chavez, G.E. Friederich, Continental-shelf sediment as a primary source of iron for coastal phytoplankton, Nature 398 (6729) (1999) 697 – 700.

562

M.S. Fantle, D.J. DePaolo / Earth and Planetary Science Letters 228 (2004) 547–562

[72] J.W.M. Wijsman, P.M.J. Herman, J.J. Middelburg, K. Soetaert, A model for early diagenetic processes in sediments of the continental shelf of the Black Sea, Estuar. Coast. Shelf Sci. 54 (3) (2002) 403 – 421. [73] I. Konig, M. Haeckel, M. Drodt, E. Suess, A.X. Trautwein, Reactive Fe(II) layers in deep-sea sediments, Geochim. Cosmochim. Acta 63 (10) (1999) 1517 – 1526. [74] J.W.M. Wijsman, J.J. Middelburg, C.H.R. Heip, Reactive iron in Black Sea sediments: implications for iron cycling, Mar. Geol. 172 (3–4) (2001) 167 – 180. [75] J. Friedrich, C. Dinkel, G. Friedl, N. Pimenov, J. Wijsman, M.T. Gomoiu, A. Cociasu, L. Popa, B. Wehrli, Benthic nutrient cycling and diagenetic pathways in the north-western Black Sea, Estuar. Coast. Shelf Sci. 54 (3) (2002) 369 – 383. [76] J.R. Schwalbach, D.S. Gorsline, Holocene sediment budgets for the basins of the California continental borderland, J. Sediment. Petrol. 55 (6) (1985) 829 – 842. [77] D.L. Inman, C.E. Nordstrom, On the tectonic and morphologic classification of coasts, J. Geol. 79 (1) (1971) 1 – 21. [78] J.M. Santana-Casiano, M. Gonzalez-Davila, F.J. Millero, The oxidation of Fe(II) in NaCl-HCO3- and seawater solutions in the presence of phthalate and salicylate ions: a kinetic model, Mar. Chem. 85 (1–2) (2004) 27 – 40. [79] J.M. Santana-Casiano, M. Gonzalez-Davila, M.J. Rodriguez, F.J. Millero, The effect of organic compounds in the oxidation kinetics of Fe(II), Mar. Chem. 70 (1–3) (2000) 211 – 222. [80] T.D. Waite, Thermodynamics of the Iron System in Seawater, in: D.R. Turner, K.A. Hunter (Eds.), The biogeochemistry of

[81]

[82]

[83]

[84]

[85]

[86]

[87]

[88]

iron in seawater, IUPAC Series on Analytical and Physical Chemistry of Environmental Systems, vol. 7, J. Wiley, Chichester, NY, 2001, pp. 291 – 342. G. Zhuang, R.A. Duce, D.R. Kester, The dissolution of atmospheric iron in surface seawater of the open ocean, J. Geophys. Res. Oceans 95 (C9) (1990) 16207 – 16216. K. Kuma, S. Nakabayashi, Y. Suzuki, K. Matsunaga, Dissolution rate and solubility of colloidal hydrous ferric-oxide in seawater, Mar. Chem. 38 (1–2) (1992) 133 – 143. K. Kuma, K. Matsunaga, Availability of colloidal ferric oxides to coastal marine-phytoplankton, Mar. Biol. 122 (1) (1995) 1 – 11. S. Nakabayashi, K. Kuma, K. Sasaoka, S. Saitoh, M. Mochizuki, N. Shiga, M. Kusakabe, Variation in iron(III) solubility and iron concentration in the northwestern North Pacific Ocean, Limnol. Oceanogr. 47 (3) (2002) 885 – 892. J.F. Wu, G.W. Luther, Spatial and temporal distribution of iron in the surface water of the northwestern Atlantic Ocean, Geochim. Cosmochim. Acta 60 (15) (1996) 2729 – 2741. J. Nishioka, S. Takeda, C.S. Wong, W.K. Johnson, Sizefractionated iron concentrations in the northeast Pacific Ocean: distribution of soluble and small colloidal iron, Mar. Chem. 74 (2–3) (2001) 157 – 179. K.S. Johnson, R.M. Gordon, K.H. Coale, What controls dissolved iron concentrations in the world ocean? Mar. Chem. 57 (3–4) (1997) 137 – 161. W. Broecker, T.H. Peng, Tracers in the Sea, Eldigio Press, Palisades, NY, 1982 (690 pp.).