Applied Geochemistry (v.38, #C)

Fractionation of U, Th, Ra and Pb from boreal forest soils by sequential extractions by Sinikka Virtanen; Kaisa Vaaramaa; Jukka Lehto (1-9).
To determine the mobility of natural radionuclides in boreal forest soil, a five-step sequential extraction procedure was carried out on soil samples taken from various depths down to 3 m on Olkiluoto Island, Finland, where there are plans to construct a spent nuclear fuel disposal repository in the bedrock. The extracted fractions studied were exchangeable, acid-soluble, reducible, oxidizable and tightly bound. It was found that the extractability of most of the radionuclides studied was dependent on the sample grain size and depth. All the elements were concentrated in the smallest grain size samples (<0.063 mm). The extraction behaviour of Th, however, did not vary with sample depth, and only about 10% of the Th was extracted by the time of the final extraction step. Stable Pb and 210Pb, as well as Ba and Ra concentrations were strongly correlated in the extractions. Radium and Ba were leached more readily than the other elements; approximately 17% of the total Ra was found in the first fraction extracted, representing exchangeable ions. Uranium was more mobile in the topsoil horizons than in the lower horizons. In the topsoil samples, an average of 51% of the extractable U was leached in the second extraction step, representing the elements soluble in weak acids, whereas only 13% of the U in the subsoil samples was extracted in this step. This is probably due to changes in soil redox conditions lower down the soil profile. The extraction behaviour of Pb and Fe also suggests the presence of more reducing conditions in the deeper soil horizons, because the percentage of extractable Pb and Fe in the oxidizable fraction increased with sample depth.

2-D reactive transport modeling of the fate of CO2 injected into a saline aquifer in the Wabamun Lake Area, Alberta, Canada by Chantsalmaa Dalkhaa; Maurice Shevalier; Michael Nightingale; Bernhard Mayer (10-23).
Carbon capture and storage (CCS) is a viable option to reduce emissions of anthropogenic CO2 into the atmosphere. One important component during the storage site selection process is the prediction of the movement and the fate of the injected CO2 in the chosen reservoir. The fate of CO2 injected into a saline aquifer of the Devonian Nisku Formation in Alberta (Canada) was simulated using the reactive transport code TOUGHREACT. A 2D radially symmetric model was developed for a 50-year CO2 injection phase followed by a 1000-year storage period. CO2 was injected into the bottom 10 m of the Nisku Formation at a rate of 1 Mton/year for 50 years. The injected supercritical CO2 spread out radially up to 3.5 km after 50 years of injection and 6.0 km after 1000 years of post-injection. After 1000 years, the injected CO2 was trapped in the Nisku Formation underneath the excellent seal of the Calmar shale predominantly as free CO2 via hydrodynamic trapping (76%) and to a lesser extent by solubility trapping (23%) in aqueous phase ( HCO 3 - ( aq ) ). Mineral trapping of injected CO2 in the Nisku aquifer was negligible. This study predicted the interactions of the fluids and rock minerals with injected CO2 in the storage Nisku aquifer, the Calmar caprock and the bottom rock Ireton shale. This information will be highly beneficial for potential future CO2 injection projects targeting the saline aquifer of the Nisku Formation in Western Canada.

Transport and fate of ammonium and its impact on uranium and other trace elements at a former uranium mill tailing site by Ziheng Miao; Hakan N. Akyol; Andrew L. McMillan; Mark L. Brusseau (24-32).
The remediation of ammonium-containing groundwater discharged from uranium mill tailing sites is a difficult problem facing the mining industry. The Monument Valley site is a former uranium mining site in the southwest US with both ammonium and nitrate contamination of groundwater. In this study, samples collected from 14 selected wells were analyzed for major cations and anions, trace elements, and isotopic composition of ammonium and nitrate. In addition, geochemical data from the U.S. Department of Energy (DOE) database were analyzed. Results showing oxic redox conditions and correspondence of isotopic compositions of ammonium and nitrate confirmed the natural attenuation of ammonium via nitrification. Moreover, it was observed that ammonium concentration within the plume area is closely related to concentrations of uranium and a series of other trace elements including chromium, selenium, vanadium, iron, and manganese. It is hypothesized that ammonium–nitrate transformation processes influence the disposition of the trace elements through mediation of redox potential, pH, and possibly aqueous complexation and solid-phase sorption. Despite the generally relatively low concentrations of trace elements present in groundwater, their transport and fate may be influenced by remediation of ammonium or nitrate at the site.

Metal release from dolomites at high partial-pressures of CO2 by Assaf Wunsch; Alexis K. Navarre-Sitchler; Joel Moore; Allison Ricko; John E. McCray (33-47).
The potential for metal release associated with CO2 leakage from underground storage formations into shallow aquifers is an important consideration in assessment of risk associated with CO2 sequestration. Metal release can be driven by acidification of groundwaters caused by dissolution of CO2 and subsequent dissociation of carbonic acid. Thus, acidity is considered one of the main drivers for water quality degradation when evaluating potential impacts of CO2 leakage. Dissolution of carbonate minerals buffers the increased acidity. Thus, it is generally thought that carbonate aquifers will be less impacted by CO2 leakage than non-carbonate aquifers due to their high buffering potential. However, dissolution of carbonate minerals can also release trace metals, often present as impurities in the carbonate crystal structure, into solution. The impact of the release of trace metals through this mechanism on water quality remains relatively unknown. In a previous study we demonstrated that calcite dissolution contributed more metal release into solution than sulfide dissolution or desorption when limestone samples were dissolved in elevated CO2 conditions. The study presented in this paper expanded our work to dolomite formations and details a thorough investigation on the role of mineral composition and mechanisms on trace element release in the presence of CO2. Detailed characterization of samples from dolomite formations demonstrated stronger associations of metal releases with dissolution of carbonate mineral phases relative to sulfide minerals or surface sorption sites. Aqueous concentrations of Sr2+, CO2+, Mn2+, Ni2+, Tl+, and Zn2+ increased when these dolomite rocks were exposed to elevated concentrations of CO2. The aqueous concentrations of these metals correlate to aqueous concentrations of Ca2+ throughout the experiments. All of the experimental evidence points to carbonate minerals as the dominant source of metals from these dolomite rocks to solution under experimental CO2 leakage conditions. Aqueous concentrations of Ca2+ and Mg2+ predicted from numerical simulation of kinetic dolomite dissolution match those observed in the experiments when the surface area is three to five orders of magnitude lower than the surface area of the samples measured by gas adsorption.

Spatial and temporal changes in groundwater salinity in South Florida by R. Ivan Blanco; G. Melodie Naja; Rosanna G. Rivero; Rene M. Price (48-58).
A combination of statistical and interpolation methods and Geographic Information System (GIS) spatial analysis was used to evaluate the spatial and temporal changes in groundwater Cl concentrations in Collier and Lee Counties (southwestern Florida), and Miami-Dade and Broward Counties (southeastern Florida), since 1985. In southwestern Florida, the average Cl concentrations in the shallow wells (0–43 m) in Collier and Lee Counties increased from 132 mg L−1 in 1985 to 230 mg L−1 in 2000. The average Cl concentrations in the deep wells (>43 m) of southwestern Florida increased from 392 mg L−1 in 1985 to 447 mg L−1 in 2000. Results also indicated a positive correlation between the mean sea level and Cl concentrations and between the mean sea level and groundwater levels for the shallow wells. Concentrations in the Biscayne Aquifer (southeastern Florida) were significantly higher than those of southwestern Florida. The average Cl concentrations increased from 159 mg L−1 in 1985 to 470 mg L−1 in 2010 for the shallow wells (<33 m) and from 1360 mg L−1 in 1985 to 2050 mg L−1 in 2010 for the deep wells (>33 m). In the Biscayne Aquifer, wells showed a positive or negative correlation between mean sea level and Cl concentrations according to their location with respect to the saltwater intrusion line. Wells located inland behind canal control structures and west of the saltwater intrusion line showed negative correlation values, whereas wells located east of the saltwater intrusion line showed positive values. Overall, the results indicated that since 1985, there was a potential decline in the available freshwater resources estimated at about 12–17% of the available drinking-quality groundwater of the southeastern study area located in the Biscayne Aquifer.

An urban stream differs from a natural stream in that it commonly contains anthropogenic nutrients from a variety of diffuse and point sources (e.g., urban runoff, industrial and municipal effluents). The Cuyahoga River as one of such stream systems receives on average 30% of water from over a dozen wastewater treatment plants (WWTPs) and exports over 300 Mg/yr of total phosphorus (TP) into Lake Erie. Municipal effluents account for at least two thirds of the TP loading and 90% of the effluent TP is in the form of soluble reactive phosphorus (SRP), i.e., the highly bioavailable form in aquatic ecosystems. Owing to its dominance of the effluent P, the Cuyahoga River may pose a disproportionately greater ecological risk to Lake Erie. Here we report results of TP and SRP as measured on water samples collected from 12 locations along the Cuyahoga River and its major tributary – Tinkers Creek under three different flow conditions in July 2007 through May 2008. Our results show that the loadings of SRP and TP increased progressively with the effluent inputs increasing from the upper basin downstream. We found the loadings of SRP and TP are not only affected by the amount of effluent P input but also regulated by stream flow regimes. Effluent P is more likely to be transformed during storm runoff events, whereas TP is more likely to be retained under low flow conditions. As a result, most of the TP loading was exported during the storm and intermediate flow conditions, whereas most of the SRP loading was delivered during the low and intermediate flow conditions. These results suggest that stream hydrology has played an important role in the loading and transport of P across the Cuyahoga River as it dictated the amount, form, and timing of P exported to Lake Erie.

The Diyadin Geothermal area, located in the eastern part of Anatolia (Turkey) where there has been recent volcanic activity, is favorable for the formation of geothermal systems. Indeed, the Diyadin geothermal system is located in an active geodynamic zone, where strike-slip faults and tensional cracks have developed due to N–S regional compression. The area is characterized by closely spaced thermal and mineralized springs, with temperatures in the range 30–64 °C, and flowrates 0.5–10 L/s. Thermal spring waters are mainly of Ca(Na)-HCO3 and Ca(Mg)-SO4 types, with high salinity, while cold groundwater is mostly of Ca(Na, Mg)-HCO3 type, with lower salinity. High contents of some minor elements in thermal waters, such as F, B, Li, Rb, Sr and Cs probably derive from enhanced water–rock interaction.Thermal water samples collected from Diyadin are far from chemical equilibrium as the waters flow upward from reservoirs towards spring vents and possibly mix with cooler waters. The temperatures of the deep geothermal reservoirs are estimated to be between 92 and 156 °C in Diyadin field, based on quartz geothermometry, while slightly lower estimates are obtained using chalcedony geothermometers. The isotopic composition of thermal water (δ18O, δ2H, δ3H) indicates their deep-circulating meteoric origin. The waters are likely to have originated from the percolation of rainwater along fractures and faults to the deep hot reservoir. Subsequent heating by conduction due to the presence of an intrusive cupola associated with the Tendurek volcano, is followed by the ascent of deep waters to the surface along faults and fractures that act as hydrothermal conduits.Modeling of the geothermal fluids indicates that the fluid is oversaturated with calcite, aragonite and dolomite, which matches travertine precipitation in the discharge area. Likewise, the fluid is oversaturated with respect to quartz, and chalcedony indicating the possibility of siliceous precipitation near the discharge areas. A conceptual hydro-geochemical model of the Diyadin thermal waters based on the isotope and chemical analytical results, has been constructed.

Evaluation of the influence of urbanization processes using mangrove and fecal markers in recent organic matter in a tropical tidal flat estuary by C.F. Grilo; R.R. Neto; M.A. Vicente; E.V.R. de Castro; R.C.L. Figueira; R.S. Carreira (82-91).
Three sediment cores (50 cm depth) were collected at three different sites from a tidal flat estuary at Passagem Channel (Vitória, Espírito Santo State-Brazil) to evaluate the influence of recent urbanization processes on the deposition of organic matter (OM) in a complex polluted tropical estuary. In addition to geochronology (by excess 210Pb), the sources of natural and anthropogenic OM to the sediments were evaluated by total organic C (TOC – 14.29 ± 8.73, 30.43 ± 14.71 and 48.70 mg g−1  ± 25.46, respectively, for P1, P2 and P3), C/N molar ratio and lipid biomarkers (sterols and terpenoids). Taraxerol (3.10 ± 4.85, 9.71 ± 3.85 and 16.10 mg gTOC−1  ± 32.48 for P1, P2 and P3, respectively) and sitosterol (1.71 ± 2.72, 2.94 ± 6.41 and 4.07 mg gTOC−1  ± 4.41 for P1, P2 and P3, respectively) were the most abundant compounds in all cores, suggesting a major contribution of terrestrially-derived OM to the study region. Coprostanol levels and selected sterol index indicated significant contamination by fecal material. The organic geochemical indicators suggest that changes of OM reflect occupation and urbanization alteration processes around the Passagem Channel over the last 70 a, mainly the conversion of mangrove forest into urban areas, bridge building and Treatment Plant Station installation.

The role of climate in the accumulation of lithium-rich brine in the Central Andes by L.V. Godfrey; L.-H. Chan; R.N. Alonso; T.K. Lowenstein; W.F. McDonough; J. Houston; J. Li; A. Bobst; T.E. Jordan (92-102).
Lithium-rich brine within the sub-surface of the Salar del Hombre Muerto (SHM) salt pan in the Andes of northwestern Argentina has a chemical and isotopic composition which is consistent with Li derived from several sources: the modern halite saturated lagoon, Li-rich salts and brines formed recently, and dissolution of halite which precipitated from ancient saline lakes. SHM lies in the closed basin that includes part of the massive Cerro Galán caldera which is drained by the Río los Patos, which is responsible for 90% of surface runoff into the salar. The low Li isotope composition, +3.4‰, of this river is consistent with significant contributions of geothermal spring water. As water drains through the volcaniclastic deposits which cover a large proportion of the basin, Li removal, as indicated by decreasing Li/Na, occurs but without significant isotope fractionation. This indicates a mechanism of surface sorption onto smectite or ferrihydrite rather than Li incorporation into octahedral structural sites of clays. These observations suggest that conditions in this high altitude desert have limited the dilution of hydrothermal spring water as well as the formation of clay minerals, which jointly have allowed the Li resource to accumulate rapidly. Changes in climate on a multi-millennial time scale, specifically in the hydrologic budget, have resulted in solute accumulation rates that have been variable through time, and decoupled Li and Na fluxes. Inflow to the salar under modern conditions has high Li/Na (7.9 × 10−3 by wt) with δ7Li indistinguishable from basement rocks (−0.3‰ to +6.4‰), while under pluvial climate conditions the Li/Na of the saline lake was 40 times lower than the modern lagoon (0.1–0.3 × 10−3 compared to 10.6–13.4 × 10−3) with slightly higher δ7Li, +6.9‰ to +12.3‰, reflecting the uptake of 6Li into secondary minerals which formed under a wetter climate.

The concentrations of dissolved cations and anions are fundamental to the water quality of groundwaters, particularily pertaining to their use as a drinking water source. Currently, assessment of total dissolved solids requires laboratory-based analysis of each ionic species, and cannot be reliably tested in the field. Here, a new method is developed that distinguishes between changes in ion concentration or water type based upon field measurement of Electric Conductivity (EC) and the concentration of only one anion. The approach operates under the assumption that all cations contribute equally to EC, whereas almost twice as much of the HCO 3 - ion is needed to produce the same effect on total solute EC as those of Cl or SO 4 2 - ions for waters at neutral pH. Therefore, many groundwater systems can be treated as two anion systems when the groundwater chemical type and solute EC relationship is assessed. Moreover, by assuming neutral charge and that the correlation between anion concentration and EC is adequately described, changes to anion concentrations will produce corresponding change to cation concentrations. A database consisting of 3555 analyses of major chemical composition from Estonian groundwaters was used to develop the correlations described here, and data for 320 samples with measured EC values together with major chemical constituents was used to verify the new method. As a result, a binary plot has been compiled which can be used: (1) for prediction of chemical composition of groundwater if analytical results of Cl or HCO 3 - from anions and measured values of EC are available, or (2) for description of temporal changes in groundwater chemistry if the time series of the same measurements are available.

The competitive adsorption of arsenate and arsenite with silicic acid at the ferrihydrite–water interface was investigated over a wide pH range using batch sorption experiments, attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy, extended X-ray absorption fine structure (EXAFS) spectroscopy, and density functional theory (DFT) modeling. Batch sorption results indicate that the adsorption of arsenate and arsenite on the 6-L ferrihydrite surface exhibits a strong pH-dependence, and the effect of pH on arsenic sorption differs between arsenate and arsenite. Arsenate adsorption decreases consistently with increasing pH; whereas arsenite adsorption initially increases with pH to a sorption maximum at pH 7–9, where after sorption decreases with further increases in pH. Results indicate that competitive adsorption between silicic acid and arsenate is negligible under the experimental conditions; whereas strong competitive adsorption was observed between silicic acid and arsenite, particularly at low and high pH. In situ, flow-through ATR-FTIR data reveal that in the absence of silicic acid, arsenate forms inner-sphere, binuclear bidentate, complexes at the ferrihydrite surface across the entire pH range. Silicic acid also forms inner-sphere complexes at ferrihydrite surfaces throughout the entire pH range probed by this study (pH 2.8–9.0). The ATR-FTIR data also reveal that silicic acid undergoes polymerization at the ferrihydrite surface under the environmentally-relevant concentrations studied (e.g., 1.0 mM). According to ATR-FTIR data, arsenate complexation mode was not affected by the presence of silicic acid. EXAFS analyses and DFT modeling confirmed that arsenate tetrahedra were bonded to Fe metal centers via binuclear bidentate complexation with average As(V)-Fe bond distance of 3.27 Å. The EXAFS data indicate that arsenite forms both mononuclear bidentate and binuclear bidentate complexes with 6-L ferrihydrite as indicated by two As(III)–Fe bond distances of ∼2.92–2.94 and 3.41–3.44 Å, respectively. The As–Fe bond distances in both arsenate and arsenite EXAFS spectra remained unchanged in the presence of Si, suggesting that whereas Si diminishes arsenite adsorption preferentially, it has a negligible effect on As–Fe bonding mechanisms.

Numerical simulations based on plans for a deep geothermal system in Basel, Switzerland are used here to understand chemical processes that occur in an initially dry granitoid reservoir during hydraulic stimulation and long-term water circulation to extract heat. An important question regarding the sustainability of such enhanced geothermal systems (EGS), is whether water–rock reactions will eventually lead to clogging of flow paths in the reservoir and thereby reduce or even completely block fluid throughput. A reactive transport model allows the main chemical reactions to be predicted and the resulting evolution of porosity to be tracked over the expected 30-year operational lifetime of the system. The simulations show that injection of surface water to stimulate fracture permeability in the monzogranite reservoir at 190 °C and 5000 m depth induces redox reactions between the oxidised surface water and the reduced wall rock. Although new calcite, chlorite, hematite and other minerals precipitate near the injection well, their volumes are low and more than compensated by those of the dissolving wall-rock minerals. Thus, during stimulation, reduction of injectivity by mineral precipitation is unlikely. During the simulated long-term operation of the system, the main mineral reactions are the hydration and albitization of plagioclase, the alteration of hornblende to an assemblage of smectites and chlorites and of primary K-feldspar to muscovite and microcline. Within a closed-system doublet, the composition of the circulated fluid changes only slightly during its repeated passage through the reservoir, as the wall rock essentially undergoes isochemical recrystallization. Even after 30 years of circulation, the calculations show that porosity is reduced by only ∼0.2%, well below the expected fracture porosity induced by stimulation. This result suggests that permeability reduction owing to water–rock interaction is unlikely to jeopardize the long-term operation of deep, granitoid-hosted EGS systems. A peculiarity at Basel is the presence of anhydrite as fracture coatings at ∼5000 m depth. Simulated exposure of the circulating fluid to anhydrite induces a stronger redox disequilibrium in the reservoir, driving dissolution of ferrous minerals and precipitation of ferric smectites, hematite and pyrite. However, even in this scenario the porosity reduction is at most 0.5%, a value which is unproblematic for sustainable fluid circulation through the reservoir.

Sources of dissolved inorganic nitrogen in a coastal lagoon adjacent to a major metropolitan area, Miami Florida (USA) by Peter K. Swart; William T. Anderson; Mark A. Altabet; Courtney Drayer; Sarah Bellmund (134-146).
Between 2006 and 2007, a study was carried out to determine the relative importance of natural and anthropogenic input of nitrogen into Biscayne Bay (South Florida, USA) using δ13C and δ15N values of algae, seagrasses, and particulate organic material, δ18O and δ15N of the NO 3 - and δ13C of the dissolved inorganic carbon. The δ15N values of all components showed a strong east to west gradient approaching more positive values (+7 to +10‰) close to the land-sea interface. The nitrogen could have emanated from the local waste water treatment plant, septic systems within the region, or nitrogen which had been affected by denitrification and leached from the local landfill, wastewater which had been injected into the Floridan aquifer and leaked back to the surface, and/or some other as yet unidentified source. The measured NO 3 - δ15N and δ18O values indicated that the dissolved nitrate originated from anthropogenic sources and was fractionated during assimilation.

In this paper we report the Sr isotope signatures, and Sr, Al and Na concentrations of 30 surface waters (lakes/ponds and rivers/creeks) and 19 soil sample extracts from the island of Bornholm (Denmark) and present a categorized 87Sr/86Sr value distribution map that may serve as a base for provenance studies, including archaeological migration and authenticity proof for particular food products. The Sr isotopic compositions of surface waters range from 87Sr/86Sr = 0.7097–0.7281 (average 0.7175 ± 0.0049; 1σ), whereas 0.1 M HNO3, 0.05 M HNO3, and 0.01 M CaCl2 soil extracts range from 87Sr/86Sr = 0.7095–0.7197 and define somewhat lower but statistically indistinguishable averages of 0.7125 ± 0.003 (1s). These compositions are lower than the values expected from the Precambrian granitoid basement (87Sr/86Sr = 0.758–0.944), and from the overlying, mainly clastic Paleozoic sediments. Combined Sr isotope composition vs. Sr, Na and Al concentration relationships of soil extracts imply that lowering of the isotopic composition of leachable Sr on Bornholm results as a consequence of significant admixture to this fraction of Sr deposited as marine salts (aerosols), and that rainwater only has a minor influence on the Sr budget of the surface waters. Positively correlated Al/Na and [1/Sr] vs. 87Sr/86Sr relationships in soil extracts and surface waters indicate that the surface run-off on Bornholm is characterized by two predominant sources, namely marine aerosols (sea salts) with high Sr and low 87Sr/86Sr values, and a source with lower [Sr] delivering radiogenic Sr to the surface waters, which we equate with Sr leached from the products of mineral weathering (soils).A feasibility study for using Sr isotopic compositions of surface waters and soil extracts as a proxy for bioavailable Sr signatures was performed with a few samples collected in the vicinity of the eleventh century AD Ndr. Grødbygård cemetery site in SW Bornholm, from where Sr isotope compositions of modern fauna samples and tooth enamel of humans buried in the cemetery have been reported. Waters and soil extracts studied herein from around this site range from 87Sr/86Sr = 0.7104–0.7166 and correspond to Sr compositions extracted from snail shells in this area which span a range of 87Sr/86Sr = 0.7095–0.7160. Some human tooth enamel is characterized by more radiogenic values (87Sr/86Sr up to 0.718) which points to a possible provenance of these humans from the granite–gneiss terrain in the north of the island and/or to immigration of these humans in their childhood from other places (for example from mainland Sweden) to Bornholm. If the total compositional range of 87Sr/86Sr = 0.709–0.718 (n  = 44) recorded in human enamel from the Ndr. Grødbygård site is considered representative for the variation of bioavailable Sr on Bornholm, then our soil leachate and surface water data entirely covers this range. We therefore propose that the combination of Sr isotope analyses of surface waters and soil leachates are an easy, fast and relatively cost efficient way to characterize a local bioavailable 87Sr/86Sr signature, and consequently propose that the overall average of 87Sr/86Sr = 0.7153 ± 0.0048 (1σ; n  = 50) can be taken as a band for bioavailable Sr fractions suitable to discriminate between local and non-local signatures in provenance studies in the field of archaeology and for food and plant authenticity control in agricultural applications.

Occurrence and mobilization of radium in fresh to saline coastal groundwater inferred from geochemical and isotopic tracers (Sr, S, O, H, Ra, Rn) by David S. Vinson; Tarik Tagma; Lhoussaine Bouchaou; Gary S. Dwyer; Nathaniel R. Warner; Avner Vengosh (161-175).
Salinization in groundwater systems can induce water–rock interaction, including the release of naturally-occurring trace elements of health significance such as radium (Ra), with possible implications for the usability of water resources in addition to the increase of dissolved solids (TDS) concentrations. In general, radium mobility is limited by chemical removal mechanisms including adsorption onto clays and/or Mn and Fe oxides, exchange processes, and coprecipitation with secondary barite. In order to examine the effect of aquifer salinity gradients on the distribution of naturally-occurring Ra in fresh to saline groundwater and the relationship to water–rock interaction and Ra removal mechanisms, two contrasting systems were investigated: the shallow unconfined coastal aquifer in Agadir (southwestern Morocco) and the confined Cretaceous (Cape Fear) and Pliocene (Yorktown) aquifers of the Atlantic Coastal Plain (North Carolina, USA). Geochemical and isotopic indicators of salinity sources (e.g. cation ratios, δ18O, δ2H, Br/Cl, δ 34 S – SO 4 2 - , δ 18 O – SO 4 2 - ) were used to identify the relative contributions of seawater and other saline waters and subsequent geochemical modification by water–rock interaction. Radium activities (224Ra, 226Ra, 228Ra), radon-222, alkaline earth metal (Mg, Ca, Sr, Ba) concentrations and ratios, and 87Sr/86Sr ratios were analyzed to identify water–rock interaction processes affecting alkaline earth metals including Ra. The Morocco coastal aquifer is generally oxic, exhibits a range of salinity and water types (Cl 163–2120 mg/L, median 932 mg/L), and exhibits Ca/Na ratios above the seawater value, typical of monovalent–divalent cation exchange (base-exchange reactions) in coastal aquifers. In contrast, the Atlantic Coastal Plain aquifers are anoxic, sulfate-reducing, cover a wider salinity range (Cl 5–9890 mg/L, median 800 mg/L) representing a transition between Na – HCO 3 - and Na–Cl waters, and exhibit Ca/Na ratios below that of modern seawater typical of reverse base-exchange reactions. Possible salinity sources in the Morocco coastal aquifer include seawater intrusion, Mesozoic evaporites, other natural saline waters, and/or wastewater, whereas the Atlantic Coastal Plain is primarily affected by old seawater present in the aquifer system. Radium activities are generally low and vary significantly within each aquifer, for example 226Ra ranges from 1.8–27.7 mBq/L in the Morocco coastal aquifer (median 8.1 mBq/L) and 1.9–214 mBq/L in the Atlantic Coastal Plain (median 18.9 mBq/L). The highest Ra activities were observed in the most saline wells sampled in the Atlantic Coastal Plain. At total dissolved solids (TDS) concentrations above an apparent threshold of ∼5000 mg/L, radium activities increase in a generally linear fashion with salinity in the Atlantic Coastal Plain, broadly comparable to previous studies indicating a threshold range of ∼3000–10,000 mg/L. At lower TDS concentrations, water–rock interaction processes that vary with local aquifer conditions appear to control Ra distribution rather than merely salinity. In the Morocco coastal aquifer, adsorption of Ra and coprecipitation with secondary barite are apparently favorable to control Ra levels in groundwater. Radium removal in the anoxic Atlantic Coastal Plain aquifers appears to be associated with adsorption and/or exchange processes, with the additional possibility of barite precipitation in the Cape Fear aquifer indicated by barite saturation. Overall, the locally-variable factors that can control Ra sources and mobility in fresh to brackish groundwater at near-neutral pH include variation in solid-phase radioactivity, redox state affecting adsorption sites, availability of competing divalent cations, and barite saturation.