Search

Published (U-Th)/He ages indicate that an oolitic Miocene channel iron deposit (CID) on Mesa J in the Hamersley province of Western Australia (∼22 °S latitude) contains three predominant generations of Fe (III) oxides – two generations of older hematite and one of younger goethite. Selective dissolution and sequential atom balance calculations were combined with chemical and oxygen isotope analyses to determine the δ18O values of hematite, goethite, and quartz in eight samples from a core drilled in that CID. For the quartz (Q), δ18OQ = +21.7(±0.9)‰. For the younger hematite (Hm1), δ18OHm1 ≥ +3.6(±0.6)‰, whereas for the older hematite (Hm2), δ18OHm2 ≤ +2.1(±0.8)‰. The oolitic CID goethite (G) crystallized in the late Miocene [7(±1) Ma] and is a solid solution represented as Fe(1−Y)AlY(CO3)XO(1−X)OH with values of δ18OG that range from −0.4‰ to +0.7‰. A mixing model for the solid solution expresses the value of δ18OG in terms of Y, X, relevant fractionation factors, and the δ18O of the endmember FeOOH (δ18Ogt). Mixing model-derived values of δ18Ogt range from −2.4‰ to −0.7‰. Paired values of δDgt and δ18Ogt in seven of eight late Miocene oolitic goethites from Mesa J indicate crystallization from meteoric waters at an average temperature of 20(±5) °C. In the late Miocene, Mesa J was in the subtropics at a paleolatitude of ∼29 °S, which invites comparison with the climate of the near-coastal community of Mingenew in Western Australia (presently at 29 °S latitude). The modern MAT at Mingenew is 20 °C – i.e., indistinguishable from the late Miocene Mesa J average of 20 °C. However, this similarity is not matched by the mean annual precipitation (MAP). The modern MAP at Mingenew is about 400 mm, but the lateritic weathering that produces abundant pedogenic Fe (III) oxides appears to require a seasonally contrasting climate with rainfall totals of at least 1300 mm/yr. In the late Miocene, southward-shifted summer season tropical cyclones might have delivered such large amounts of rain (with relatively low δD and δ18O values). If so, the δDgt and δ18Ogt of the oolitic goethites of Mesa J support the idea of a rainfall-driven, possibly microbially mediated, summer season bias in the timing of crystallization. Thermal buffering in the regolith can reconcile summer-dominated goethite crystallization with a temperature that would be analytically indistinguishable from the MAT. Comparison of the δ18Ogt values of the Mesa J goethites with published δ18O values of goethites of similar age from two other sites in the Hamersley province suggest a possible increase in δ18O of late Miocene rainfall with increasing distance from the coast.

The channel iron deposits (CID) of the Hamersley Province in Western Australia are dominated by pedogenic goethite/hematite-rich ooids and pisoids that were transported to, and deposited in, the meandering channels of Miocene rivers. Information about the Miocene weathering environments that produced the Fe(III) oxides is archived in the mole fraction (X) and δ13C of the Fe(CO3)OH component in solid solution in oolitic CID goethite (α-FeOOH). Values of X and δ13C were measured for 12 oolitic goethite samples from different depths in two cores drilled in CID of the Robe Formation of Mesa J. The weighted-average plateau values of X ranged from 0.0098 to 0.0334, which suggest ambient CO2 concentrations that ranged from ∼50,000 ppm V to perhaps as much as ∼200,000 ppm V at the time of goethite crystallization. In a vadose zone characterized by in situ production of CO2 with steady-state Fickian diffusive transport of the gas, such concentrations would correspond to modeled soil respiration rates (Q) ranging from about 10 to 30 mmol/m2/h. Values for Q of about 10 mmol/m2/h are reported for soils in modern tropical forests with MAP ≥ ∼2000 mm. However, model-derived values of Q that exceed 15 mmol/m2/h are larger than observed in modern systems. This could indicate that some of the CID goethites crystallized in conditions that were phreatic or near phreatic rather than vadose. The δ13C values of the Fe(CO3)OH component in these 12 CID samples ranged from −24.0‰ to −22.3‰, which are among the most negative measured to date. If they reflect steady-state diffusive transport of CO2 in vadose environments, the soil CO2 would have been derived from a source with δ13C values that ranged from ∼−31‰ to −29‰. If, on the other hand, the goethites crystallized in a nearly phreatic environment that was moderately acidic, the inferred δ13C of the ancient CO2 source would have been about −27.6‰ to −25.8‰. In either case, the δ13C values point to in situ oxidation of C3 organic matter as the predominant source of the ambient CO2. The Fe(III) oxides in the CID ooids suggest crystallization in aerobic environments. However, even in aerobic environments, many microbial species can reduce the Fe3+ in oxides to relatively soluble Fe2+ and may have facilitated progressive Fe enrichment during multiple cycles of Fe(III) oxide dissolution and recrystallization. At the same time, microbially mediated oxidation of organic matter could have produced the high concentrations of soil CO2 with the very negative δ13C values recorded in the Fe(CO3)OH component in oolitic goethite. More frequent summer storms in the Miocene, may have been a significant factor in forming and eroding these soil systems and in concentrating large volumes of oolitic Fe(III) oxides in the local river systems to form channel iron deposits. However, published (U-Th)/He ages indicate that the oolitic CID goethites of Mesa J became closed systems after ∼7 Ma, which suggests a change in local climate and/or conditions of burial at about that time in the Miocene.

Nineteen (U-Th)/He ages were determined for eight samples from a core drilled in an ore-grade channel iron deposit (CID) of the Robe Pisolite (Robe Formation) of Mesa J in Western Australia. With one exception, uncorrected ages of the analyzed aliquots range from 6.7(±0.4) Ma to 30.2(±3.1) Ma, while molar ratios of Th/U range from 0.42 to 5.06. The exception is an aliquot with an apparent age of 2.7 Ma and Th/U of 5.70. A three-component mixing model involving one generation of goethite and two generations of hematite suggests that the age of crystallization of the oolitic goethites is ∼7(±1) Ma. If so, the goethites have effectively been closed systems for ∼7 million years and should preserve a stable hydrogen isotope record of late Miocene rainfall in the vicinity of Mesa J. Cenozoic movement of the Australian continent had placed Mesa J and environs in the subtropics at a paleolatitude of about 29 °S during the late Miocene. Al-adjusted δD values of oolitic goethite in the eight CID samples range from −153‰ to −146‰ and imply that the δD of the late Miocene meteoric waters ranged from −61‰ to −53‰, with an average of −56‰. These relatively negative δD values might indicate that near-coastal, late Miocene rain was derived primarily from summer-season tropical cyclones with storm tracks that extended into the subtropics of western Australia. The postulated late Miocene tropical cyclones would have occurred more often and/or exhibited greater intensity at a paleolatitude of 29 °S than is the case for modern sites at approximately 30 °S on the west coast of Australia (e.g., Perth). Higher fluxes of meteoric water in the Miocene summers would have facilitated dissolution and removal of BIF-sourced silica with concomitant enrichment in oxidized Fe. Moreover, wetter late Miocene summers could have promoted multiple cycles of microbially mediated dissolution and recrystallization of Fe(III) oxides in the aerobic systems. The oolitic textures may be indicative of such recycling. However, the oolitic goethites of Mesa J were closed systems after ∼7 Ma. Therefore, the climate in the vicinity of Mesa J seems to have changed in the late Miocene to conditions that did not favor widespread recycling of Fe (III) oxides—perhaps changing from seasonally wet to the modern dry climate.

The initial ∼60% of an isothermal vacuum dehydration of goethite can commonly be approximated by first order kinetics. Also, natural goethites contain small amounts of an Fe(CO3)OH component in apparent solid solution. The 18O/16O of CO2 evolved from the Fe(CO3)OH during isothermal vacuum dehydrations is related to the 18O/16O of the goethite by an apparent fractionation factor (αapp) that is, in turn, correlated with a first order rate constant, |m|. A kinetic exchange model predicts that αapp should decrease as |m| increases for a range of |m| that corresponds to relatively slow rates of dehydration. This pattern has been observed in published results. In contrast, for rapid rates of dehydration, αapp is predicted to increase with increasing |m|. Isothermal vacuum dehydrations of two natural goethites had unusually large values of |m| and provided serendipitous tests of this rapid-rate prediction. For these experiments, the measured values of αapp were consistent with patterns of variation predicted by the model. This allowed an estimate of the activation energy (E2) of a model parameter, K2, which is the rate constant for oxygen isotope exchange between CO2 and H2O during the solid-state goethite to hematite phase transition. The estimated value of E2 is only ∼9 kJ/mol. Heterogeneous catalysis tends to decrease the activation energies of gas reactions. Consequently, the inferred value of E2 suggests that goethite and/or hematite catalyze oxygen isotope exchange between CO2 and H2O during the solid-state phase change. Yield, δ13C, and δ18O values are routinely measured for increments of CO2 evolved from the Fe(CO3)OH component during isothermal vacuum dehydration of goethite. Model-predicted values of αapp can be combined with plateau δ18O values of the evolved CO2 to estimate the δ18O of the goethite with a less than optimal, but potentially useful, precision of about ±0.8‰. Therefore, a single analytical procedure (incremental dehydration) applied to one aliquot of a sample could provide not only δ13C and mole fractions (X) of the Fe(CO3)OH component, but also hydrogen yield, δD, and the approximate δ18O value of the goethite. Recovery of multiple types of geochemical data from a single aliquot would be particularly useful if the amount of sample available for analysis were limited. Also, the method could be used to estimate the δ18O value of goethite in mixtures of minerals not amenable to selective dissolution – e.g., goethite admixed with hematite.

Measurements of the fractional abundance (m) of high temperature nonstoichiometric (HTN) hydrogen in goethite are shown to predict the reduction in O2 yield associated with removal of HTN oxygen by routine pre-treatment of a sample with BrF5 at 22 °C. The predicted relationship between smaller O2 yields and m is linear and defines an “oxygen yield limit” (OYL), which is introduced for use in oxygen isotope analyses of goethite. The OYL sets lower limits for O2 yields that would be consistent with: (1) removal of all HTN oxygen during pre-treatment and (2) complete recovery of the goethite structural oxygen during subsequent reaction with BrF5 at 450 °C. Measured O2 yields that are less than those predicted by the OYL are anomalously low and indicate probable partial loss of structural oxygen during pre-treatment at 22 °C. After such partial loss, measured δ18O values of the residue are more negative than the actual δ18O of the total structural oxygen. Therefore, the OYL can be used to distinguish between robust and suspect δ18O analyses of goethite. This OYL criterion was applied to some newly analyzed synthetic goethites in which Al substitutes for Fe up to mole fractions (XAl) of ∼0.13. As expected, the robust δ18O values indicate that 1000 ln 18α (where, 18α = [18O/16Omineral]/[18O/16Owater]) increases with increasing XAl, but comparison of these results with the predictions of a published solid solution model suggests that the relationship between 1000 ln 18α and XAl is sensitive to very high pH and possibly to the concentration of aqueous Cl− during synthesis of goethite. The cause of the apparent sensitivity to Cl− remains unknown. The various relationships are: equation(A) T=42°C,pH=1.5–2.5,[Cl-]≈0.16M:1000ln18α=18±5XAl+1.8±0.3 equation(B) low-pH model,[Cl-]=0.00M:1000ln18α=13(±2)XAl+4.1(±0.3) equation(C) T=22°C,pH=14,[Cl-]=0.00M:1000ln18α=5(±3)XAl+1.6(±0.3) equation(D) low-pH model,[Cl-]=0.00M:1000ln18α=13(±2)XAl+6.5(±0.3) However, for natural freshwater goethites, the low-pH, no-Cl, solid solution model (e.g., isothermal equations B and D) best predicts the oxygen isotope effects associated with substitution of Al for Fe. Thus, for goethites from freshwater systems with acidic to circum-neutral pH, the following expression can be used to calculate end member goethite (FeOOH) δ18O values from the measured values of δ18O and XAl: δ18OFeOOH=δ18Omeasured-13(±2)XAlδ18OFeOOH=δ18Omeasured-13(±2)XAl Because the slope of the preceding expression appears to be relatively insensitive to temperature, the equation should be applicable over the range of temperatures at which goethite commonly crystallizes in natural systems (0 °C ⩽ T ⩽ ∼70 °C). However, the value of δ18OFeOOH is sensitive to temperature via the temperature dependence of 1000 ln 18α for pure goethite at acidic to circum-neutral pH in freshwaters as follows (T in degrees Kelvin): 1000ln18α=1.66×106T2-12.6 The results of this study affirm the need to correct measured δ18O values of goethite for the effects of substituent Al to obtain useful paleoenvironmental information, and they support the existing model for doing so.

Goethite (Ax-2) from Axel Heiberg Island (∼80°N) on the margin of the Arctic Ocean is the dominant mineral in a sample of “petrified” Eocene wood, but U, Th, and He measurements suggest that the goethite (α-FeOOH) crystallized in the latest Miocene/Pliocene (ca. 5.5 to 2.8 Ma). Measured δD and δ18O values of Ax-2 are −221 (±6)‰ and −9.6 (±0.5)‰, respectively. The inferred δD and δ18O values of the ancient water were about −139‰ and −18.6‰, respectively, with a calculated temperature of crystallization of 3 (±5)°C, which compares with the modern summer (J-J-A) temperature of 3 °C and contrasts with a modern MAT of −19 °C. Published results from various biological proxies on nearby Ellesmere Island indicate a Pliocene (∼4 Ma) MAT of either −6 or −0.4 °C and corresponding seasonal amplitudes of about 18 or 13 °C. A conductive heat flow model suggests that a temperature of 3 °C could represent goethite crystallization at depths of ∼100–200 cm (for MAT = −6 °C) or ∼250–450 cm (for MAT = −0.4 °C) over seasonally restricted intervals of time. The δ18O value of the Ax-2 water (−18.6‰) is more positive than the modern J-J-A precipitation (−22‰). In combination, the paleotemperatures and δ18O values of ancient waters (from Ax-2 and published results from three Eocene or Pliocene proxy sites on Axel Heiberg and Ellesmere Islands) are consistent with a warm season bias in those isotopic proxies. The results are also consistent with higher proportions of J-J-A precipitation in the annual total. If so, this emphasizes the importance of seasonality at high latitudes even in times of warmer global climates, and suggests that the Arctic hydrologic cycle, as expressed in the seasonal distribution and isotopic composition of precipitation (perhaps modified by a warmer Arctic Ocean), differed from modern. The δ13C value of the Fe(CO3)OH component in the Ax-2 goethite is +6.6‰, which is much more positive than expected if crystallizing goethite incorporated CO2 derived primarily from oxidation of relict Eocene wood with δ13C values of about −24‰. This apparent paradox may be resolved if the goethite is a product of oxidation of 13C-rich siderite, which had previously replaced wood in an Eocene methanogenic burial environment. Thus, the goethite retains a carbon isotope “memory” of a diagenetic Eocene event, but a δD and δ18O record of the latest Miocene/Pliocene Arctic climate.

The isotopic composition of land snail shells was analyzed to investigate environmental changes in the eastern Canary Islands (28–29°N) over the last ~ 50 ka. Shell δ13C values range from −8.9‰ to 3.8‰. At various times during the glacial interval (~ 15 to ~ 50 ka), moving average shell δ13C values were 3‰ higher than today, suggesting a larger proportion of C4 plants at those periods. Shell δ18O values range from −1.9‰ to 4.5‰, with moving average δ18O values exhibiting a noisy but long-term increase from 0.1‰ at ~ 50 ka to 1.6–1.8‰ during the LGM (~ 15–22 ka). Subsequently, the moving average δ18O values range from 0.0‰ at ~ 12 ka to 0.9‰ at present. Calculations using a published snail flux balance model for δ18O, constrained by regional temperatures and ocean δ18O values, suggest that relative humidity at the times of snail activity fluctuated but exhibited a long-term decline over the last ~ 50 ka, eventually resulting in the current semiarid conditions of the eastern Canary Islands (consistent with the aridification process in the nearby Sahara). Thus, low-latitude oceanic island land snail shells may be isotopic archives of glacial to interglacial and tropical/subtropical environmental change.

From July to November 2009, concentrations of CO2 in 78 samples of ambient air collected in 18 different interior spaces on a university campus in Dallas, Texas (USA) ranged from 386 to 1980 ppm. Corresponding δ13C values varied from −8.9‰ to −19.4‰. The CO2 from 22 samples of outdoor air (also collected on campus) had a more limited range of concentrations from 385 to 447 ppm (avg. = 408 ppm), while δ13C values varied from −10.1‰ to −8.4‰ (avg.=-9.0‰). In contrast to ambient indoor and outdoor air, the concentrations of CO2 exhaled by 38 different individuals ranged from 38,300 to 76,200 ppm (avg. = 55,100 ppm), while δ13C values ranged from −24.8‰ to −17.7‰ (avg. = −21.8‰). The residence times of the total air in the interior spaces of this study appear to have been on the order of 10 min with relatively rapid approaches (∼30 min) to steady-state concentrations of ambient CO2 gas. Collectively, the δ13C values of the indoor CO2 samples were linearly correlated with the reciprocal of CO2 concentration, exhibiting an intercept of −21.8‰, with r2 = 0.99 and p The average concentration of outdoor Dallas atmospheric CO2 was ∼17 ppm higher than the average of CO2 concentrations measured on the same campus 10 a ago. In addition, Dallas outdoor CO2 concentrations at both times were higher than the contemporaneous global atmospheric CO2 concentrations. This observation, plus the fact that the increase of ∼17 ppm in the average concentration of Dallas outdoor CO2 was comparable to the global increase of ∼18 ppm over the same 10-a interval, is consistent with a significant role for urban CO2 “factories” in the global atmospheric CO2 budget.

Kaolinite, gibbsite and quartz are the dominant minerals in samples collected from two outcrops of a Cenomanian (∼95 Ma) laterite in southwestern Minnesota. A combination of measured yields and isotope ratios permitted mass balance calculations of the δ D and δ 18 O values of the kaolinite in these samples. These calculations yielded kaolinite δ D values of about −73‰ and δ 18 O values of about +18.7‰. The δ D and δ 18 O values appear to preserve information on the ancient weathering system. If formed in hydrogen and oxygen isotope equilibrium with water characterized by the global meteoric water line (GMWL), the kaolinite δ D and δ 18 O values indicate a crystallization temperature of 22 (±5) °C. A nominal paleotemperature of 22 °C implies a δ 18 O value for the corresponding water of −6.3‰. The combination of temperature and meteoric water δ 18 O values is consistent with relatively intense rainfall at that mid-paleolatitude location (∼40°N) on the eastern shore of the North American Western Interior Seaway. The inferred Cenomanian paleosol temperature of ∼22 °C is in general accord with published mid-Cretaceous continental mean annual temperatures (MAT) estimated from leaf margin analyses of fossil plants. When compared with results from a published GCM-based Cenomanian climate simulation which specifies a latitudinal sea surface temperature (SST) gradient that was either near modern or smaller-than-modern, the kaolinite paleotemperature of 22 °C is closer to the GCM-predicted MAT for a smaller equator-to-pole temperature difference in the mid-Cretaceous. Moreover, the warm, kaolinite-derived, mid-paleolatitude temperature of 22 °C is associated with proxy estimates of high concentrations of atmospheric CO 2 in the Cenomanian. The overall similarity of proxy and model results suggests that the general features of Cenomanian continental climate in that North American locale are probably being revealed.

An Early Eocene Oxisol in the Ione Formation of California formed in a coastal continental weathering environment at a paleolatitude of ∼38°N. The dominant minerals in the Oxisol are goethite, quartz, and kaolinite. Material balance calculations were applied to new measurements of chemical composition, D/H, and 18 O/ 16 O ratios of Oxisol samples to determine the δD (−150 ± 3‰) and δ 18 O (−2.4 ± 0.3‰) values of the goethite (α-FeOOH). These data, in combination with the global meteoric water line (MWL), yielded an isotopic temperature of 21(±4) °C. The nominal value of 21 °C contrasts with the modern mean annual temperature (MAT) of 16 °C in that area. The warmer temperature is consistent with formation of the goethite during the Early Eocene climatic optimum. The isotopic composition of the goethite and a temperature of 21 °C imply ancient water with a δD value of −61(±4)‰ and a δ 18 O value of −8.9(±0.5)‰. This Early Eocene δ 18 O (or δD) value is more negative than values in the range of isotopic scatter observed for modern global precipitation at sites with a MAT of 21 °C. At times of warm global climates, the location of a near-surface atmospheric isotherm would generally shift relative to its location under modern climatic conditions. A simple Rayleigh-type condensation model indicates that, if one “follows the isotherm”, the associated scatter in δD and δ 18 O of precipitation in very warm global climates should shift (for a given isotherm) to more negative values that may be detectable in proxy records. The isotopic results from the goethite of the Early Eocene Oxisol appear to add to evidence in support of this idea.

Twelve goethite samples with different degrees of substitution of Al for Fe were synthesized at 22°C to 48°C and pH values of 1.5 to 14 under closed system conditions and used to study the effects of Al substitution on the hydrogen isotopic fractionation between goethite and its ambient water. The syntheses followed two pathways: 1) Fe 3+ hydrolysis in high pH aqueous solutions; 2) oxidation of Fe 2+ to Fe 3+ in mid to low pH solutions. XRD and SEM analyses indicated that, irrespective of temperature and pH, goethite was the predominant product of the syntheses in all of the experiments (with degrees of Al substitution as high as ~13 mole percent). “High temperature nonstoichiometric” (HTN) water is present in all of the samples and rapidly exchanges D/H with ambient vapor at room temperature. Uncertainties in the value of the apparent D/H fractionation factor ( e-v) between HTN water and ambient exchange water at 22°C lead to significant uncertainties in determinations of the D values of structural hydrogen ( Ds) in goethites which contain high proportions of HTN water. As determined for the samples of this study, e-v has a nominal value of 0.942 (±0.02). Ds values determined using an e-v value of 0.942 indicate that Al substitution increases the D value of structural hydrogen in goethite by about 1.4 (±0.4) ‰ for each increase in Al of 1 mole %. This dependence on Al is of the same sign as, but somewhat larger in magnitude than, the effect of Al predicted by a published model (~0.7‰ per mole % Al). The overall uncertainties in the current results suggest that an increase of ~1‰ per mole % Al, as adopted by previous studies, may be a reasonable estimate with which to adjust Ds values of natural goethites to those of the pure FeOOH endmember and could be valid for degrees of Al substitution of up to at least 15 mole %. These synthesis experiments also yield a hydrogen isotopic fractionation factor ( D G-W) between pure goethite ( -FeOOH) and liquid water of 0.900 (±0.006), which is analytically indistinguishable from the published value of 0.905 (±0.004). Thus, use of an D G-W value

Goethite synthesis experiments indicate that, in addition to temperature, pH can affect the measured value of the 18O/16O fractionation factor between goethite and water (αG–W). A simple model was developed which expresses αG–W in terms of kinetic parameters associated with the growth of goethite from aqueous solution. The model predicts that, at a particular temperature, the range of pH over which αG–W changes as pH changes is expected to be comparatively small (∼3 pH “units”) relative to the range of pH values over which goethite can crystallize (pH from ∼1 to 14). Outside the range of sensitivity to pH, αG–W is predicted to be effectively constant (for constant temperature) at either a low-pH αG–W value or a high-pH αG–W value. It also indicates that the values of αG–W at high pH will be disequilibrium values. Values of αG–W for goethite crystallized at low pH may approach, but probably do not attain, equilibrium values. For goethite synthesized at values of pH from ∼1 to 2, data from two different laboratories define the following equation for the temperature dependence of 1000 ln αG–W (T in degrees Kelvin) equation(IV) 1000lnαG–W=1.66×106T2-12.6 Over the range of temperatures from 0 to 120°C, values of 1000 ln αG–W from Eq. (IV) differ by ⩽0.1‰from those of a published equation [Yapp C.J., 1990. Oxygen isotopes in iron (III) oxides. 1. Mineral–water fractionation factors. Chem. Geol.85, 329–335]. Therefore, interpretations of data from natural goethites using the older equation are not changed by use of Eq. (IV). Data from a synthetic goethite suggest that the temperature dependence of 1000 ln αG–W at low pH as expressed in Eq. (IV) may be valid for values of pH up to at least 6. This result and the model prediction of an insensitivity of αG–W to pH over a larger range of pH values could explain the observation that Eq. (IV) yields values of αG–W which mimic most 18O/16O fractionations measured to date in natural goethites.

A Late Paleocene (∼60 Ma BP) lateritic soil from Northern Ireland (the Antrim paleosol, herein referred to as Nire) contains coexisting goethite, gibbsite, phyllosilicate, and hematite. The Fe(III) oxides exhibit pisolitic and Liesegang-type morphologies that are mutually exclusive in hand specimens. X-ray diffraction (XRD) measurements of Al substituted for Fe in goethite indicate two populations: (1) low-Al, Liesegang-type goethites (∼0 mol% Al) and (2) high-Al, pisolitic goethites (∼9 to ∼24 mol% Al). Selective dissolution and incremental vacuum dehydration-decarbonation were used to determine the concentration and δ13C values of CO2 occluded in the respective structures of the goethites and gibbsites in this complex mixture of Nire lateritic minerals. The Fe(CO3)OH component in the high-Al goethites appears to retain a proxy carbon isotopic record of vadose zone CO2 in the ancient soil. The δ13C values of CO2 occluded in coexisting goethites and gibbsites indicate that these minerals did not form in equilibrium with the same environmental CO2. The measured mole fractions (X) of Fe(CO3)OH in the high-Al goethites range from 0.0059 (±0.0005) to 0.0077 (±0.0006) and correspond to soil CO2 concentrations of ∼28,000 to ∼37,000 ppmV. The average values of X and δ13C for the four high-Al goethites are 0.0067 ± 0.0007 and −20.1 ± 0.5‰, respectively. The δ13C value of the organic matter undergoing oxidation in this midlatitude (∼55°N) Late Paleocene soil appears to have been ∼ −28.2‰. Taken together, these data indicate an atmospheric CO2 concentration of ∼2400 ppmV (± ∼1200 ppmV) at ∼60 Ma BP. The inferred high concentration of atmospheric CO2 would have been coincident with the warm global climate of the Late Paleocene and is consistent with the idea that CO2 plays an important role in climate variation.

Incremental vacuum dehydration-decarbonation experiments were performed at 190°C on chemically “cleaned” aliquots of a gibbsite-dominated, Eocene-age bauxite sample with evolution of CO 2 and H 2 O. “Plateau” F (CO 2 /H 2 O ratios) and δ 13 C values of the CO 2 derived from gibbsite were attained over the dehydration interval, X v (H 2 ) = 0.16 to 0.67 (i.e., 16 to 67% breakdown of gibbsite). The plateau value of F for gibbsite was 0.0043 ± 0.0003, while the corresponding δ 13 C value of evolved CO 2 was −16.0‰±0.4‰. Additional experiments on chemically cleaned aliquots included (1) treatment with a solution of 0.3M Na-Citrate + 0.1M Na-Dithionite and (2) an exchange experiment with 0.1 bar of 13 C-depleted CO 2 (−46‰) at 105°C for 64.5 h. Neither of these additional treatments resulted in a measurable perturbation of plateau values of F or δ 13 C for CO 2 evolved from gibbsite during dehydroxylation. These results support published work on Holocene samples which suggested that CO 2 occluded in gibbsite may preserve information on δ 13 C values of CO 2 in ancient terrestrial systems. The plateau values of F observed in the Eocene gibbsite indicate that it may be possible to experimentally calibrate a relationship between the concentration of CO 2 occluded in gibbsite and CO 2 in the environment at the time of crystallization. Such a calibration would significantly enhance the value of gibbsite as a source of information on ancient oxidized carbon systems.

13C/12C and 18O/16O ratios of aragonite shells of modern land snails from the southern Great Plains of North America were measured for samples from twelve localities in a narrow east–west corridor that extended from the Flint Hills in North Central Oklahoma to the foothills of the Sangre de Cristo Mountains in Northern New Mexico, USA. Across the study area, shell δ18O values (PDB scale) ranged from −4.1‰ to 1.2‰, while δ13C values ranged from −13.2‰ to 0.0‰. δ18O values of the shell aragonite were predicted with a published, steady state, evaporative flux balance model. The predicted values differed (with one exception) by less than 1‰ from locality averages of measured δ18O values. This similarity suggests that relative humidity at the time of snail activity is an important control on the δ18O values of the aragonite and emphasizes the seasonal nature of the climatic information preserved in the shells. Correlated δ13C values of coexisting Vallonia and Gastrocopta suggest similar feeding habits and imply that these genera can provide information on variations in southern Great Plains plant ecology. Although there is considerable scatter, multispecies, transect average δ13C values of the modern aragonite shells are related to variations in the type of photosynthesis (i.e., C3, C4) in the local plant communities. The results of this study emphasize the desirability of obtaining isotope ratios representing averages of many shells in a locale to reduce possible biases associated with local variations among individuals, species, etc., and thus better represent the “neighborhood” scale temporal and/or spatial environmental variations of interest in studies of modern and ancient systems.

A simple flux balance model with a diffusive, evaporative boundary layer indicates that the time constant (characteristic time) for approach to oxygen isotope steady state in the body fluid of land snails is ∼19 min or less. These comparatively short times support an assumption that the snail’s aragonitic shell is commonly precipitated from a body fluid that is at, or near, isotopic steady state. The model indicates that the steady-state δ18O value of snail shell carbonate depends upon the temperature, relative humidity, δ18O of the input liquid water, and δ18O of ambient water vapor. Model shell δ18O values were calculated for the warm, wet months corresponding to times of snail activity at some European sites. Linear regression of these predicted values against published, measured values yielded the expression: δ18Ocalc = 0.93(±0.13) δ18Omeas −0.9(±0.2), with r2 = 0.65. As indicated by the value of r2, there is scatter in the relationship, but the slope and intercept are close to one and zero, respectively, which lends credence to the model. Therefore, temporal or spatial changes recorded in the δ18O values of land snail shells appear to be selectively seasonal—commonly the warm, wet months—and include the effects of relative humidity. For carbon, the time constant for approach to isotopic steady state in the bicarbonate dissolved in the body fluid of land snails is predicted to be ∼16 min or less. New and published δ13C measurements of aragonite shell and associated organic matter exhibit an overall correlation, but with considerable scatter. As noted by previous workers, 13C-rich dietary “limestone” may account for some of the scatter. Additional scatter, according to the model presented herein, could arise from changes in the proportion of total oxidized carbon that is expelled by the snail as bicarbonate dissolved in body fluid (i.e., effects of relative changes in metabolic rates). These results affirm the need for caution in the interpretation of δ13C values of land snail aragonite shells solely in terms of dietary proportions of C3 and C4 plants.

Pedogenic goethites in each of two Early Permian paleosols appear to record mixing of two isotopically distinct CO2 components—atmospheric CO2 and CO2 from in situ oxidation of organic matter. The δ13C values measured for the Fe(CO3)OH component in solid solution in these Permian goethites are −13.5‰ for the Lower Leonardian (∼283 Ma BP) paleosol (MCGoeth) and −13.9‰ for the Upper Leonardian (∼270 Ma BP) paleosol (SAP). These goethites contain the most 13C-rich Fe(CO3)OH measured to date for pedogenic goethites crystallized in soils exhibiting mixing of the two aforementioned CO2 components. δ13C measured for 43 organic matter samples in the Lower Leonardian (Waggoner Ranch Fm.) has an average value of −20.3 ± 1.1‰ (1s). The average value yields a calculated Early Permian atmospheric P co 2 value of about 1 × PAL, but the scatter in the measured δ13C values of organic matter permits a calculated maximum P co 2 of 11 × PAL (PAL = present atmospheric level). Measured values of the mole fraction of Fe(CO3)OH in MCGoeth and SAP correspond to soil CO2 concentrations in the Early Permian paleosol profiles of 54,000 and 50,000 ppmV, respectively. Such high soil CO2 concentrations are similar to modern soils in warm, wet environments. The average δ13C values of pedogenic calcite from 9 paleosol profiles stratigraphically associated with MCGoeth (Waggoner Ranch Fm.) range from −6.5‰ to −4.4‰, with a mean δ13C value for all profiles of −5.4‰. Thus, the value of Δ13C between the pedogenic calcite data set and MCGoeth is 8.1 (±0.9)‰, which is in reasonable accord with the value of 7.7‰ expected if atmospheric P co 2 and organic matter δ13C values were the same for both paleosol types. Furthermore, the atmospheric P co 2 calculated for the Early Permian from the average measured carbon isotopic compositions of the paleosol calcite and organic matter is also analytically indistinguishable from 1 × PAL, with a maximum calculated atmospheric P co 2 (permitted by one standard deviation of the organic matter δ13C value) of ∼5 × PAL. If, however, measured average δ13C values of the plant organic matter are more positive than the original soil organic matter as a result of diagenetic loss of 13C-depleted, labile organic compounds, calculated Permian atmospheric P co 2 using these 13C-enriched organic values would underestimate the actual atmospheric P co 2 using either goethite or calcite. This is the first stratigraphically constrained, intrabasinal study to compare ancient atmospheric CO2 concentrations calculated from pedogenic goethite and calcite. These results demonstrate that the two different proxies record the same information about atmospheric CO2. The Fe(CO3)OH component in pedogenic goethite from a Triassic paleosol in Utah is significantly enriched in 13C relative to Fe(CO3)OH in goethites from soils in which there are mixtures of two isotopic CO2 components. Field-relationships and the δ13C value (−1.9‰) of the Triassic goethite indicate that this ancient paleosol profile experienced mixing of three isotopically distinct CO2 components at the time of goethite crystallization. The three components were probably atmospheric CO2, CO2 from in situ oxidation of organic matter and CO2 from in situ dissolution of preexisting calcite. Although mixing of three isotopically distinct CO2 components, as recorded by Fe(CO3)OH in goethite, has been described in modern soil, this is the first example from a documented paleosol. Its preservation affirms the need for careful, case-by-case assessment of ancient paleosols to establish that goethite in any particular soil is likely to be a valid proxy of atmospheric P co 2.

Measured mole fractions (X) and δ13C values of the Fe(CO3)OH component in pedogenic goethite from a mid-latitude Oxisol of Early Eocene age (≈52 Ma B.P.) range from 0.0014 to 0.0064 and −20.1 to −15.4‰, respectively. These values of X imply that concentrations of CO2 gas in the paleosol were ≈7400 to ≈34,000 ppm. δ13C and 1/X are correlated and define a linear, soil-CO2 diffusive mixing line with a positive slope. Such positive slopes are characteristic of mixing of two isotopically distinct CO2 endmembers (atmospheric CO2 and CO2 from oxidation of soil organic matter). From the intercept of the mixing line, it is calculated that the δ 13C value of organic matter in the ancient soil was ≈−28.0‰. The magnitude of the slope implies an Early Eocene atmospheric CO2 concentration of ≈2700 ppm. A simple model for forest soils suggests that a “canopy effect” may cause atmospheric CO2 concentrations deduced from pedogenic minerals to underestimate the actual concentrations of atmospheric CO2. If a significant forest canopy were present at the time of formation of pedogenic goethite in the Ione Fm, the concentration of 2700 ppm calculated for atmospheric CO2 could be slightly low, but the underestimate is expected to be

“Plateau” δ 18 O values of CO 2 that evolved from the Fe(CO 3 )OH component during isothermal vacuum dehydrations (200–230 °C) of 18 natural goethites range from 8.2 to 28.1‰. In contrast, the measured δ 18 O values of the goethite structural oxygen range from −11.3 to 1.7‰. The results of this study indicate that the apparent oxygen isotope fractionation factor ( 18 α app ) between plateau CO 2 and initial goethite is systematically related to the rate of isothermal vacuum dehydration. The nonlinear correlation and the magnitudes of the 18 α app values are predicted by a relatively simple mass balance model with the following assumptions: (1) the rate of isothermal vacuum dehydration of goethite (for the interval from 0 to ∼60 to 80% loss of structural hydroxyl hydrogen) can be reasonably well represented by first-order kinetics and (2) isotopic exchange between evolving H 2 O vapor and solid occurs only in successive, local transition states. The generally good correspondence between the model predictions and the experimental data seems to validate these assumptions. Thus, the 18 O/ 16 O ratios of the evolved CO 2 can act as probes into the transient processes operating at the molecular level during the solid-state goethite-to-hematite phase transition. For example, the activation energy for the rate constant associated with the transition state, oxygen isotopic exchange between solid and H 2 O vapor, is tentatively estimated as 28 ± 11 KJ/mol. Such knowledge may be of consequence in understanding the significance of 18 O/ 16 O ratios in hematites from some natural environments (e.g., Mars?). Kinetic data and δ 18 O values of CO 2 are routinely obtained in the course of measurements of the abundance and δ 13 C values of the Fe(CO 3 )OH in goethite. The observed correlation between 18 α app and dehydration rates suggests that plateau δ 18 O values of evolved CO 2 may provide complementary estimates of the δ 18 O values of total goethite structural oxygen (O, OH, CO 2 ) with an overall precision of about ±1‰. However, because of isotopic exchange during the dehydration process, δ 18 O values of the evolved CO 2 do not reflect the original δ 18 O values of the CO 2 that was occluded as Fe(CO 3 )OH in goethite.

The concentrations and δ13C values of atmospheric CO2 were measured in ∼150 air samples collected at 8 sites in the Dallas metropolitan area over the period August 1998 to December 1999. Measured concentrations (C) of atmospheric CO2 ranged from 369 to 475 ppm, while the δ13C values ranged from –12.0 to –8.1‰. These values contrast with a “global” concentration at the time of this study of approximately 367 ppm and a corresponding δ13C value of about –8.0‰. δ13C was linearly correlated with 1/C for samples collected at heights of ∼2 m at 3 sites adjacent to streets with significant automobile traffic. Extrapolation of this two-component mixing line to 1/C=0 yielded a δ13C value of about –27‰ for the CO2 input—i.e., the same as that of gasoline. A simple box model, incorporating photosynthesis, respiration, and anthropogenic addition of CO2, indicates that differences between downwind and upwind concentration-weighted δ13C values ( Δ [ C ∗ δ13C]) of atmospheric CO2 may be linearly correlated with downwind and upwind differences in concentration ( C d ∗ − C u ∗ ), where C ∗ is reported as mol/m3. The model predicts that measurable effects of photosynthetic withdrawal of atmospheric CO2 are manifested by data arrays with slopes more positive than about –16. This effect of photosynthesis is evident in a linear array of “warm weather”, Dallas atmospheric CO2 data (slope of –12.7‰). Collectively, the data for all 8 sites exhibited considerable scatter about binary mixing lines that depict the addition of CO2 from combustion of natural gas and gasoline. However, when model slopes (m) were calculated for binary mixing between a “background” atmospheric CO2 and each individual sample, it was found that, in general, m increases with decreasing temperature. The effects of photosynthesis and respiration complicate this relationship, but the overall pattern suggests that, as temperature decreases, the proportion of anthropogenic CO2 derived from combustion of natural gas increases. This increase appears to reflect increased use of natural gas for home heating, etc., in cooler weather. Therefore, seasonally changing patterns of fossil fuel use are detectable in the atmospheric CO2 of this urban environment.

Paired D/H and 18O/16O analyses of natural goethites range from −224‰ to −93‰ for δD and from −15.5‰ to +2.8‰ for δ18O. The isotopic systematics of these goethites seem to be most closely mimicked by the goethite–water D/H and 18O/16O fractionation curves of Yapp 1987 , Yapp 1990 ). Most of these goethites are of unknown age, but the large range of their isotopic compositions illustrates the possible paleoenvironmental significance of “surface domains” calculated for hydroxyl minerals. A modern surface domain (MSD) for goethite in coordinates of δD and δ18O was calculated from the following information: (a) scatter in the modern relationship between surface temperatures and δ18O of meteoric water (Rozanski et al., 1993) ; (b) average annual surface air temperatures from 0°C to 30°C for the meteoric water sample sites; (c) the meteoric water line (MWL) of Craig (1961) ; and (d) the goethite–water isotopic fractionation factors of Yapp 1987 , Yapp 1990 ). Goethites that had formed in continental environments in approximate thermal equilibrium with the Earth’s modern atmosphere would be expected to lie within the MSD, and 16 of 31 goethites do so. The remaining 15 are “outliers.” Of the 16 MSD samples, 12 (including the three samples from active soil or bog environments) are “concordant.” The calculated isotopic temperatures of the concordant goethites range from 6°C to 27°C and are the same (within analytical uncertainty) as the modern average surface air temperatures at the respective sites. Calculated temperatures of the four “discordant” MSD samples range from 7°C to 25°C and are higher than corresponding modern surface temperatures. A “warm Earth surface domain” (WESD) was calculated with an assumed temperature of 35°C for the tropics. Eight of the 15 outlier samples plot within the portion of the WESD that does not overlap the MSD. Thus, these eight goethites may have formed at times when the Earth’s global climate was warmer than at present. Temperatures calculated from the eight WESD samples range from 20°C to 33°C and suggest lower latitudinal temperature gradients as predicted by climate models of a warmer Earth. Calculated temperatures of the seven samples that lie outside both the MSD and the WESD (non-sd samples) range from 19°C to 69°C. The non-sd status of these seven samples suggests that they formed in environments dominated by subsurface heat sources. Surface domains can be calculated for any hydroxyl mineral for which both the hydrogen and oxygen isotope mineral–water fractionation factors are known. The concept of surface domains provides a criterion for assessing the likelihood that a particular sample records information on ancient climatic temperature.

δ13C values of the Fe(CO3)OH component in pedogenic goethites from a young soil developed on the Eocene Weches formation in east Texas increase from approximately −13 or −14‰ at depths of 31–64 cm to values around −6 to −4‰ at depths greater than 122 cm. This spatial distribution of δ13C values suggests that dissolution of precursor fossiliferous marine calcite (still present at deeper levels in the soil) has contributed significantly to the isotopic budget of CO2 in this soil. A local isotopic material balance was calculated for the soil CO2 at each sample depth using a calcite δ13C value of −1‰, an organic matter δ13C value of −25‰, and the measured δ13C value of the Fe(CO3)OH component in each sample. Although there is presently no calcite in the upper 120 cm of the profile, the calculated apparent contribution of CO2 from a calcite source ranges from 16% to 45% at these shallow depths. Below 120 cm, dissolution of calcite appears to contribute more than 50% of the CO2 in the soil gas. Similar results might be expected in other wet, goethite-bearing soils that contain relict calcite and thus have not achieved the highly leached characteristics of laterites (such as those in the Amazon basin). Models of the soil CO2 budget in such systems may need to consider both oxidation of organic matter and dissolution of carbonate minerals as in situ sources of carbon isotope variation in CO2. The variation with depth of the δ13C values of the Fe(CO3)OH component suggests an ongoing process of goethite dissolution and reprecipitation in the active, aerobic soil zone. If so, the extremely low solubility of goethite in oxidizing environments suggests that this dissolution process is probably biologically mediated. Dissolution and reprecipitation of goethite in an aerobic soil would favor the recording of steady-state soil CO2 δ13C patterns. Preservation of such information in ancient soils would probably depend upon burial and consequent removal from the biologically active soil zone.

Material balance calculations were applied to the results of selective dissolution, X-ray diffraction, chemical and isotopic analyses to determine the δ18O values of oolitic goethites (α-FeOOH). Well-preserved ooids from low paleolatitude (≤35°) ironstones of Late Ordovician, Early Jurassic, Early Middle Jurassic, and Late Middle Jurassic ages have goethite δ18O values that range from −2.3 to +1.7 ‰. Late Cretaceous goethite from a high paleolatitude (55°N) deposit has a δ18O value of −5.2 ‰. The measured δ18O values of the ancient goethites are significantly different from the δ18O values expected for goethites which would form in the modern environments at each site. In contrast, there is good agreement between measured and hypothetical δ18O values of goethites from a variety of young deposits. This suggests that the ancient goethites have not isotopically reequilibrated with modern waters and supports an assumption that these oolitic goethites have preserved δ18O values reflecting ancient environments of formation. All of the oolitic goethite δ18O values analyzed for this work are consistent with formation in continental settings. A comparison with the isotopic systematics of modern, globally distributed meteoric waters indicates that the four oolitic goethites from low paleolatitudes probably formed in environments with average rainfall that exceeded 1200 mm/yr. The high paleolatitude oolitic ironstone formed near the western margin of the North American Late Cretaceous seaway. Few modern sites from the global IAEA network appear to be reasonable thermal and isotopic analogs for the Late Cretaceous environment of this deposit. Therefore, it is possible that the ancient seaway modified the Late Cretaceous hydrologic cycle to produce a local meteoric water with a combination of thermal and isotopic signatures that is not common at present. δ18O values of mineral pairs from the Late Ordovician and Early Jurassic deposits yield calculated paleotemperatures of about 23°C and 32°C (±4°C), respectively. Both epochs appear to have been characterized by atmospheric PCO2 values that were as much as 16–18 times higher than modern. The association of a comparatively low tropical temperature (and an ice sheet at high paleolatitude) with high atmospheric PCO2 might have been a result of lower solar luminosity in the Late Ordovician.

Hydrogen and oxygen isotope ratios have been measured in goethites from a lateritic regolith in the eastern Amazon Basin of Brazil and a young bog iron deposit in New Jersey, USA. The presence of exchangeable hydrogen and admixed minerals required the use of material-balance calculations to determine relevant δD and δ18O values. For the Brazilian goethite, δD is −121‰ and δ18O of −1.2‰ (adjusted for Al content). These values yield a calculated temperature of formation of 24 ± 3°C, which is in good agreement with the modern average annual temperature in that locale. Goethite from the New Jersey bog has a δD of −134‰ and δ18O of −0.9‰. Its calculated temperature of formation (23 ± 3°C) resembles summer temperatures in southeastern New Jersey. Bacterially mediated rates of precipitation of ferric hydroxide appear to be higher in the bog during the summer months (Crerar et al., 1979). Thus, the temperature calculated for the New Jersey goethite could indicate that the conversion of ferric hydroxide to goethite also occurs predominantly in the summer. δ13C values of the Fe(CO3)OH component in goethite (−21.7‰ for Brazil and −19.3‰ for New Jersey) reflect the δ13C values of the associated organic carbon (−28.4‰ for Brazil and −26.3‰ for the New Jersey). Values of PCO2 in the regolith (0.063 bar) and bog (0.039 bar) were calculated from the measured abundances of the Fe(CO3)OH component. The values of δ13C and PCO2 obtained from the goethites suggest that the ambient CO2 in these systems originated primarily from oxidized organic matter. In the low-pH, high-PCO2 waters characteristic of bogs and laterites, isotopic equilibrium may be closely approached during the formation of goethite from ferric hydroxide (ferrihydrite), even if the initial precipitation of the ferric hydroxide precursor is bacterially mediated. Therefore, meaningful paleotemperatures might commonly be calculated from stable isotope ratios of nonexchangeable oxygen and hydrogen in geothites from such environments.

Concentrations of the Fe(CO 3 )OH component in goethites from Phanerozoic oolitic ironstones appear to record information on the partial pressures of soil CO 2 in ancient subaerial weathering environments. Application of a simple steady-state, one-dimensional, Fickian diffusion model to ancient goethite-bearing soils suggests that it may be possible to calculate lower limits for the partial pressure of oxygen in the Earth's atmosphere by using both the inferred soil CO 2 partial pressure and estimates of the partial pressure of atmospheric CO 2 . Extant data from colitic goethites indicate that the atmospheric P O 2 value was no lower than about 13% of the present atmospheric level (PAL) in the Late Ordovician. This value affirms existing evidence for abundant molecular oxygen in the Earth's atmosphere in the Early Paleozoic, i.e., before the widespread advent of vascular plants. Extensive colonization of the continents by vascular plants in the Devonian was associated with calculated atmospheric oxygen partial pressures that were no lower than about 39% of PAL at 360 Ma BP. For Early Jurassic to Late Cretaceous samples, the calculated lower limit of atmospheric oxygen ranged from about 20 to 25% of the present value. It remains to be established whether or not there is a systematic relationship between calculated minimum P o 2 values (or soil respiration rates) and the actual partial pressure of atmospheric oxygen.

Abundance and carbon isotope data from an Fe(CO3)OH component in apparent solid solution in oolitic goethites have been used to infer ancient atmospheric CO2 pressures. A test of the validity of these estimates might be comparisons of the carbon isotope compositions of Fe(CO3)OH in oolitic goethites with time-equivalent pedogenic calcites. Temporal trends of the oolitic goethite and pedogenic calcite δ13C values are generally similar, but time-equivalent samples from each of these two groups are not common in the existing data. To facilitate discussion of the concept, comparisons were made of available goethite and calcite samples even though ages of the compared samples in each pair were not identical. In four out of the five comparisons, Fe(CO3)OH abundance and δ13C data were combined with pedogenic calcite δ13C data to calculate physically reasonable soil CO2 concentrations for the ancient calcitic soids. This suggests that the compared oolitic goethite and pedogenic calcite systems were responding to the same global scale phenomenon (i.e., atmospheric CO2). Atmospheric PCO2 as determined from the goethites in these four “well-behaved” cases ranged from values indistinguishable from modern (within analytical uncertainty) to values up to approximately 16 time modern (modern atmospheric PCO2 was taken to be 10−3.5 atm). One interpretation of the fifth, “anomalous”, comparison is that atmospheric CO2 levels increased from about 3 times modern to about 18 times modern from the Triassic into the Early Jurassic. This inferred value for the PCO2 of the Early Jurassic atmosphere is not uniquely constrained by the existing data and needs to be substantiated. However, even considerably lower Early Jurassic atmospheric PCO2 values of 6 to 9 times modern (i.e., 1/3 to 1/2 of the estimated value of 18 times modern) would still indicate significant differences between the global carbon cycles then and now. These results highlight the need for more research on the behavior of the atmosphere during and after the Triassic-Jurassic transition.

Hydrogen isotope exchange experiments between goethite and water vapor were performed at 22°C after vacuum “outgassing” at temperatures ranging from 100 to 140°C. The results for five natural and one synthetic goethite can be most simply interpreted in terms of two hydrogen reservoirs in the mineral. One reservoir is stoichiometric, structural hydroxyl hydrogen (FeOOH) which does not seem to exchange at 22°C on the timescales of these experiments. The other reservoir, as indicated by chemical analyses, appears to be nonstoichiometric hydrogen which persists in the sample even after outgassing at 100 to 140°C. This “high temperature” nonstoichiometric (HTN) hydrogen rapidly exchanges isotopes with water vapor at 22°C and seems to have reached, or closely approached, exchange equilibrium after only 15 to 17 hours. The isotopic fractionation factor (αe−v) between the HTN hydrogen and water vapor is estimated to be about 0.996 at 22°C. The exchange experiments indicate that the D/H ratio of stoichiometric hydrogen is not changed by the procedure for outgassing at 100 to 140°C. The amount of HTN hydrogen ranges from nearly zero to about 20% of the total hydrogen among these samples. When present in higher proportions, exchange experiments of the type described here may be necessary to determine the δD value of structural OH in goethite. However, outgassing at about 100°C in vacuum prior to dehydration at 850°C appears to be sufficient for measurement of δD values of structural OH in many natural goethites.

Carbon isotope exchange experiments and data from natural samples in a state of isotopic disequilibrium indicate that the apparent Fe(CO3)OH component in natural goethites is a closed system. These results support the solid solution model for Fe(CO3)OH in goethite. The carbon isotope geochemistry of the oolitic Neda Formation ironstone at occurrences in Wisconsin and Iowa is consistent with goethite formation in a Late Ordovician subaerial weathering environment. δ13C values of the Fe(CO3)OH component in Neda Formation goethite indicate that organic matter was being oxidized to produce CO2 in the ancient weathering profile. The δ13C value of this organic matter was about −27‰. At depths greater than about 20 cm, the partial pressure of CO2 in the Late Ordovician weathering profile was 5.6 times larger than the PCO2 of the Earth's atmosphere at that time. This high “soil” CO2 partial pressure and its origin in the oxidation of organic matter suggest that there was substantial biological activity on continental land surfaces prior to the widespread colonization by vascular plants. It indicates a possible role for biological activity in the chemical weathering of continents in the early Paleozoic.

Geological evidence suggests that the goethite-dominated, oolitic ironstone of the Upper Ordovician Neda Formation in the north central United States was subaerially weathered in the Late Ordovician. The oxygen isotope ratios of the goethites in the [open quotes]hard[close quotes] and [open quotes]soft[close quotes] ores of this deposit are rather uniform with a [delta][sup 18]O value of about [minus]1.0[per thousand]. This [delta][sup 18]O value indicates the presence of meteoric water at the time of goethite formation and supports the idea that the goethites in hard and soft ores were formed in the same chemical weathering environment 440 million years ago. The [delta][sup 18]O value of accessory phosphate in some soft ore goethite-rich ooids was measured and combined with the goethite [delta][sup 18]O value to calculate a temperature of formation of about 23[degrees]C. The corresponding [delta][sup 18]O value of the ancient water was calculated to have been [minus]7.3[per thousand]. The relatively [open quotes]cool[close quotes] temperature calculated for the low-altitude, tropical site occupied by the Neda ironstone in the Late Ordovician is consistent with published evidence from fossil marine invertebrates for global cooling at that time. The [delta][sup 18]O value of [minus]7.3[per thousand] for the ancient tropical precipitation could indicate intense rainfall ofmore » a seasonal (perhaps monsoonal) character. 27 refs., 7 figs., 3 tabs.« less

Treatment of iron (III) oxides with hot 5 M NaOH solution can selectively dissolve silicate impurities (Kampf and Schwertmann, 1982). In the current work 5M NaOH solutions enriched in {sup 18}O (+670 {per thousand}) were employed to determine if this method could be used to purify goethites for {delta}{sup 18}O analysis. These experiments suggest that the structural stoichiometric oxygen in well-crystallized goethites does not exchange with hot 5 M NaOH solution. Therefore, this selective dissolution method appears to be isotopically viable. {sup 18}O- normal' 5 M NaOH treatments were applied to goethite-dominated ooids of the Upper Ordovician Neda Fm. ironstone. While not completely removed by successive NaOH treatments, the impurities were incrementally dissolved in constant elemental proportions (within analytical error). Consequently, the {delta}{sup 18}O value of the endmember goethite could be determined by material balance calculations. This goethite {delta}{sup 18} value is {minus}1.0{per thousand} for all analyzed samples of Neda Fm. ooids, including those from occurrences about 200 km apart. The spatial uniformity of the oolitic goethite {delta}{sup 18}O values suggest uniform conditions of goethite formation. The conditions might have been those of a low latitude continental weathering environment.

Quartz (SiO 2 ) and hematite (Fe 2 O 3 ) possess considerably different gram formula weights and different stoichiometric coefficients for their oxygen. These differences mean that the numerical values of SiO 2 and Fe 2 O 3 compositions when reported as weight percent are not the same as the numerical values of X (O), i.e. oxygen in the mineral as a mole percent of the total oxygen in a binary quartz-hematite system. Consequently, end-member quartz and hematite δ 18 O-values which are determined from linear extrapolations of δ 18 O vs. weight percent Fe 2 O 3 data from binary mixtures can be in error by as much as several per mil. Extrapolation to end-member δ 18 O-values in binary quartz-hematite systems should always be performed using mole fraction of oxygen as the compositional variable. The hematite-water fractionation curve of Yapp and the quartz-water curve of Knauth and Epstein were used together with the kinetic isotope exchange model of Criss et al. to constrain the depositional environment of the Late Proterozoic quartz-hematite banded iron formation deposits of Urucum, Brazil. The depositional temperatures permitted by the model assumptions employed here range from 0° to ∼ 35°C (or 54°C). The permitted range of depositional water δ 18 O-values is from −6.6‰ (or −8.9‰) to 0.0‰. While the range of permitted depositional conditions is rather large, it does encompass the possible cool temperatures and brackish or fresh waters which are indicated by sedimentologic evidence in these deposits.

Samples of synthetic and natural goethites were subjected to isothermal dehydration under both closed- and open-systein conditions at various teinperatures ranging from 160 to 300°C. The oxygen isotope ratios of the dehydration product (hematite) were systematically different for open- and closedsystein dehydration. In every instance, open-systein dehydration resulted in 18O enrichments of the product hematite relative to the starting goethite. For all but a few cases, closed-system dehydration produced an 18O depletion in the residual mineral relative to the starting goethite. These oxygen isotope effects indicate that significant mineral-vapor isotope exchange occurred during the solid-state, closed-system transformation of goethite to∗∗ hematite. Calculated values of the apparent closed-system oxygen isotope fractionation factor between hematite and the corresponding liquid water suggest that equilibrium isotope exchange may have been approached in those samples for which the extent of closed-system dehydration was at least 95%. The rapidity of the mineral-vapor isotope exchange associated with the solid-state goethite to hematite transformation is indicated by the fact that the dehydration times (for those samples apparently approaching equilibrium) ranged from only 24 to 68 hours. These results suggest that the oxygen isotope ratios of many hematites in low temperature geological systems reflect the environmental conditions associated with their formation from some hydrous precursor. Information about the 18 O 16 O ratio in the hydrous precursor itself is not likely to be directly preserved in the hematite

CHEMICAL weathering of rocks by CO2—an important process in the carbon cycle—may have been affected by the appearance of vascular plants during the Silurian period (438–408 Myr ago), because of the enhanced efficacy of these plants in promoting weathering1. The magnitude of this effect is uncertain, however2–4. One key issue is the concentration of CO2 in soils, generated by microbial respiration; this is itself in part a function of soil productivity. We showed recently5 that the CO2 partial pressure in a Late Ordovician soil was 5.6 times that in the atmosphere; Keller and Wood6 have shown, however, that soil CO2 concentrations can be high even for low levels of microbial respiration, as a result of slow diffusion of respired CO2 out of the soil. Here we present an analysis of the Fe(CO3)OH component of goethite (FeOOH) in an Upper Ordovician oolitic ironstone which underwent tropical chemical weathering 440 Myr ago7,8. These data allow us to extract an estimate of the CO2 flux from the soil profile. The high rate that we obtain implies that the productivity of pre-vascular biota was similar to that in modern soils. Thus, attempts to model atmospheric CO2 concentrations over Phanerozoic 9–12 time need not assume that pronounced productivity changes accompanied the development of vascular plants.

THE role of changing atmospheric CO2 concentrations in controlling global temperature can be investigated by examining variations in both CO2 and climate preserved in the Earth's geological record. A model of the Earth's carbon cycle over the past 570 Myr suggests that, compared to its present value, the partial pressure of CO2 ( ) may have been an order of magnitude higher in the early Palaeozoic, and about 4–6 times higher in the middle Mesozoic1,2. Cerling3 used carbon isotope ratios in soil carbonate minerals to constrain atmospheric m portions of the Mesozoic and Cenozoic eras. Studies of the common mineral goethite (α-FeOOH) have shown that it contains small quantities of a carbonate component (Fe(CO3)OH), the concentration and carbon isotope content of which preserves a record of ambient at the time of formation4–7. Here we present data for goethites from an ironstone in the Upper Ordovician Neda Formation (Wisconsin, USA)8, which suggest that 440 Myr ago atmospheric was ~ 16 times higher than today. However, this enhanced level of atmospheric CO2 does not seem to have been accompanied by unusually warm temperatures in the tropics, and in fact may have been contemporaneous with high-latitude continental glaciation on Gondwanaland9,10.

Isotopic climate records in tree rings were obtained by the δD analyses of the hydrogen in cellulose nitrate extracted from tree rings in a Scots pine from Loch Affric, Scotland, and from a bristlecone pine from the White Mountains, California. This method of analysis measures δD values of only the isotopically non-exchangeable hydrogen of the cellulose in wood and thus eliminates serious complications in the δD record caused by the chemical heterogeneity of wood and by the isotopic exchangeability of some of its hydrogen. The average δD values of the two pines are markedly different, reflecting the contrasting climates of the two areas. The bristlecone δD record contains a 22-year periodicity perhaps recording a possible 20- to 22-year periodicity of drought conditions in the Great Plains of North America. There is no such significant periodicity in the δD record of the Scots pine. The long-term δD trends in the two pines, as represented by 40-year running averages of the δD data, correlate linearly over the time period 1841–1970 A.D., which is the total growth period of the Scots pine. The long-term δD trend of the Scots pine correlates well with the 1841–1970 winter temperatures of Edinburgh, Scotland. The long-term δD trend of the bristlecone pine, which extends over the time period 970–1970 A.D., correlates qualitatively with long-term climatic trends estimated by LaMarche from tree ring width data and by Lamb from many sources of climatic data. δD analyses were also made on early and late woods from the same annual ring. The δD values of these woods differ to various degrees and the sign of the difference can also vary.

Goethite samples analyzed for this study have a δD range from −202 to −98 per mil with a corresponding δD range of associated waters from about −110 to +7 per mil. Goethites with the most positive δD values are from marine environments. Goethite-water equilibrium D H fractionation factors measured in this laboratory at 100°C and 145°C and estimated for sedimentary temperatures from data on natural samples have values of about 0.900 at all three temperatures. These data suggest that goethite δD values may be a direct, temperature-independent measure of the δD values of the waters with which the goethite last equilibrated. Most of the goethite δD values in this study reflect the δD values of the modern waters in the locales of origin. Substitution of Mn and/or Al for Fe in the goethite structure may affect the mineral-water fractionation factor. Manganese appears to decrease the value of α. The effect of Al substitution on α has not yet been measured, but preliminary arguments suggest that increasing Al content could increase α values.

δ13C and δ18O analyses have been performed on the aragonite shells of a variety of modern land snails from a number of different geographic and climatic locales. The δ18O values of the waters assumed to be in isotopic equilibrium with the shell carbonate were calculated. These calculated δ18O values are more positive than the δ18O values of the average meteoric waters in the locales in which the snails lived. The 18O enrichment appears to be linearly related to the reciprocal of the local relative humidity, which is consistent with the notion that these ambient waters have undergone isotopic steady-state evaporation. Measurements of the δ18O values of ancient land snail shells from the excavation of Sudden Shelter (42SV6) at Ivie Creek, Utah, suggest that the climate at this site was probably warmer and/or drier around 7100–7800 BP than at present.

The dehydration of a natural goethite to hematite is accompanied by a systematic hydrogen isotope fractionation. Closed system dehydration at, and below, 250°C results in a significantly greater degree of isotopic fractionation than does open system dehydration. This relationship is apparently reversed at 300°C. Both processes produce a progressive decrease in the D H ratio of the mineral hydrogen with increasing degree of dehydration. At temperatures of 160°C to 250°C the closed system mineralvapor fractionation factor is independent of temperature, while above 250°C, it varies strongly with temperature. The mineral-vapor fractionation factor associated with open system dehydration appears to be independent of temperature over the interval 160°C to 300°C. The closed system D H fractionation suggests that natural goethite undergoing dehydration in the presence of water can isotopically exchange with that water. CO 2 loss from goethite during dehydration is correlated with the loss of H 2 O. The CO 3 is thought to be present in carbonates which exist as impurities in the goethite. Loss of both H 2 O and CO 2 appears to be diffusion-controlled.

Dehydration-decarbonation, isotopic and mass-balance data suggest that a solid solution between goethite and iron (III) carbonate may exist in natural samples. The following reaction is suggested as a possibly suitable representation of equilibrium in such a system: FeOOH + CO2⇌Fe(CO3)OH The mole fractions of this proposed iron(III) carbonate in the goethites studied thus far range from ∼0.00066 to ∼0.013. With such low concentrations it has been assumed, as a first approximation, that the proposed solid solutions can be considered ideal. A tentative calibration of CO2 isobars for the above reaction has been attempted using measured trapped CO2 and isotopic temperature data for a natural goethite from a documented marine environment, as well as information obtained from the literature. The calibration indicates a range of CO2 partial pressures of ∼0.003-0.1 bar for the natural non-marine goethites of this study. This range of PCO2 for the goethites compares to a calculated range of PCO2 of ∼0.00032-0.08 bar in modern groundwaters in limestones. Thus, although the calibration of the proposed goethite-iron(III) carbonate PCO2 indicator is tentative, it appears to be yielding results which are geologically reasonable. Much more work (including long-term laboratory syntheses of goethite under controlled PCO2 conditions) needs to be done to substantiate the solid-solution model proposed here. However, the preliminary ideas and results presented in this paper suggest that the iron oxyhydroxide, goethite, may be a source of quantitative information on CO2 partial pressures in low-temperature geologic systems.

δD and δ^(13)C values have been measured for the unexchangeable hydrogen and the total carbon of cellulose extracted from 40 North American ^(14)C-dated trees that range in age from 9500 to 22,000 years B.P. Meteoric waters which precipitated over ice-free regions of North America in the interval 14,000–22,000 B.P. had more positive δD values than corresponding modern waters by an average of 19‰. Lower ocean temperatures and smaller temperature gradients than exist at present between ocean and ice-free North America are indicated for the late Wisconsin glacial maximum. This is compatible with warmer winters and cooler summers for this glacial period. The δD value of the North American ice sheet during the Late Wisconsin maximum was approximately −100‰ as determined from the inferred δD values of the waters of proglacial lakes Agassiz and Whittlesey. From this figure the increase in δ^(18)O of the oceans during the glacial maximum can be calculated to have been +0.8‰. At the point where they began to move over the ice, air masses supplying moisture to the North American ice sheet contained a little more than 50% of their original moisture content, which is a much greater percentage than exists in air masses supplying the modern Greenland and Antarctic ice sheets. This relatively vapor-rich air coupled with lower summer temperatures, which reduced ablation, probably contributed to the maintenance and growth of the ice sheet. The transition from glacial to interglacial conditions on North America was rapid and occurred within a 2000–3000-year interval. However, the transition may not have been synchronous over North America. A 40-year δD record in a spruce branch from the Two Creeks (Wisconsin) forest (∼11,800 B.P.) shows large variations which suggest an unusual hydrologic environment in the area of the tree. Cellulose δ^(13)C values range between −20.8 and −25.9‰, but do not correlate with δD variations for the samples analyzed in this work. Thus, climatic significance of δ^(13)C variations cannot be resolved from these data.

The enzymatic decarboxylation of glutamic acid shows a carbon isotope effect k12/k13 = 1.018 at 37 degree C, pH 4.7. In D2O under otherwise identical conditions, k12/k13 = 1.009. Under the same conditions solvent isotope effects are Vmax H2O/Vmax D2O = 5.0 and (Vmax/Km)H2O/(Vmax/Km)D2O = 2.6. With the assumption that the carbon isotope effect on the decarboxylation step is in the usual range (1.05--1.07), it is possible to derive relative rates and solvent isotope effects for all steps in the enzyme mechanism. Substrate binding in approximately 2-fold weaker in H2O than in D2O, probably because of the desolvation which accompanies binding of the substrate to the enzyme. A proton inventory analysis of the reaction shows that the Schiff base interchange has a large solvent isotope effect composed of relatively small contributions from at least four separate sites. A conformation change probably accompanies this step. The decarboxylation step shows a solvent isotope effect of approximately 2. Schiff base interchange and decarboxylation are both partially rate determining. The pH dependence of the isotope effects indicates that the initial step in the reaction can occur by way of two different pathways.

Carbon, as a possibly “indigenous” minor element impurity in goethites, appears to exist in two principal forms: 1. (1) CO2 trapped in the mineral structure and 2. (2) organic matter. With one exception the total carbon content of the samples in this study ranges between 0.2 and 2.0 weight percent (reported as CO2). Total carbon δ13C values have a range from −27.2 to −8.1 per mil. Values of the mole fraction of trapped CO2 in the goethite total carbon as determined for seven samples vary from 0.11 to 0.50. Much of the variation of total carbon δ13C values in these samples can be attributed to different trapped CO2/organic matter ratios in the different goethites. By analogy with speleothems and soil carbonates, δ13C values of trapped CO2 could indicate whether the environment of goethite formation was relatively open or closed to mixing with CO2 from the Earth's atmosphere. In goethite samples subjected to H2O2 treatment, calculated organic matter δ13C values range from about −35 to −24 per mil. These values probably reflect a predominance of organic matter derived from C3 plants in this particular sample population. 13 C 12 C ratios in organic matter associated with goethite could be a source of information on the Earth's ancient biosphere.

The method of Epstein et al. (1976) for analysis of D/H ratios of cellulose carbon-bound hydrogen has been modified. This modified “renitration” method yields δD values which are in agreement with those obtained by the sodium chlorite delignification method. Comparison of results obtained by the renitration method with the published results of Epstein et al. (1976) indicate some differences in the δD values of individual samples. However, the overall plant-water δD relationship determined by Epstein et al., is not greatly changed upon redetermination by the renitration method. Additional data from a variety of plants representing a wide geographical range reveal that relative humidity is an important variable in determining the δD value of cellulose C-H hydrogen on this inter-regional scale. The role of relative humidity can be reasonably explained by a leaf water model that assumes an isotopic steady-state during transpiration. These results reaffirm the conclusion of Epstein et al., that the δD variations of the source water are the dominant control of the δD variations of cellulose C-H hydrogen from naturally grown plants. Thus, there is an expectation that these cellulose δD variations can have climate significance.

Meteoric waters in the vicinity of Albuquerque, New Mexico, U.S.A. exhibit a wide range of δD-values. Local precipitation, sampled on a monthly basis, over a 2-yr. period had δD-values ranging between − 158 and − 6‰ with the most negative values in the winter months and the most positive in the summer. Waters of the Rio Grande at Albuquerque ranged in δD from − 103 to − 79‰ over the same 2-yr. period. However, for about half the sample period, Rio Grande δD-values were within 2‰ of a “baseline” value of − 92‰. The most negative Rio Grande δD-values were produced by the contribution of spring meltwater, while runoff from summer rain produced positive shifts in the river-water δD-values. The principal components of local groundwater recharge have significantly different δD-values: (1) average Rio Grande water with δD of − 92‰; and (2) recharge from the Sandia and Manzano Mountains with δD-values of − 83 to − 70‰. Groundwater δD-values in the local aquifer (the Upper Tertiary Santa Fe Formation) vary from − 104 to − 77‰ over a horizontal distance of 20 km. The δD data suggest that there are distinct domains of groundwater within the Santa Fe Formation which retain their isotopic identities over relatively long time scales. Isotopic differences between these domains seem to reflect the differences in their respective local sources of recharge. The most negative δD-values appear to characterize deeper groundwater in the domain of Rio Grande recharge and may reflect climatic conditions different from those at present.