Gaboury Benoit

Chemical processes at the sediment-water interface*

Abstract In natural waters, the sediment-water interface is the site where gradients in physical, chemical and biological properties are the greatest. Both chemical and microbiological transformation processes are responsible for cycling elements between water and sediments. This paper discusses the various chemical transformations that take place during early diagenesis in sediments, and which are fueled by supply rates of organic carbon and electron acceptors. In particular, our current knowledge of the cycling of the electron acceptors O, N, Mn, Fe and S is assessed, and the important role of transport reactions is described. The elemental fluxes across the sediment-water interface can be described from first principles only if the coupling of physical, chemical and biological processes is better understood. Large gradients in chemical potentials of many chemical species across this interface often lead to steep gradients in proton and electron activities. The redox sequence commonly observed in time and space also leads to changes in the partitioning of solutes between carrier and solution phases. Such processes can cause rapid back-diffusion of released species to the overlying water column or their removal onto secondary carrier phases within the sediments. Physical transport systems such as resuspension or bioturbation of particles control the physical boundary conditions for chemical reactions immediately above or below the sediment-water interface. Because of the complex coupling of the different chemical, physical and biological processes, chemical reactions within this region must be considered an integral part of a three-dimensional network of interactions. Hence, the apparent coexistence of chemical species considered ‘incompatible’ by thermodynamic models can be a consequence of the three-dimensional nature of redox gradients. The microbiological, chemical and physical interactions of dissolved, colloidal and particulate organic carbon, and iron and manganese oxides play a crucial role in the early diagenetic reactions at the sediment-water interface. Their effects on diagenetic reactions, and on trace elements in solution and adsorbed to particles, are discussed. Furthermore, we emphasize the way in which hydrodynamics can control elemental fluxes in environments with high carbon rain rates.

Partitioning of Cu, Pb, Ag, Zn, Fe, Al, and Mn between filter-retained particles, colloids, and solution in six Texas estuaries

There are few methodical studies of processes controlling trace metal behavior in estuaries, where river water gradually mixes with seawater leading to systematic changes in ionic strength, pH, DOC, nutrient concentrations, and alkalinity. We report here data on Cu, Zn, Pb, and Ag behavior in six Texas estuaries using state-of-the-art ultra-clean techniques. In addition, Fe, Al, and Mn data are presented for one of Texass major estuaries, Galveston Bay. It was found that suspended matter concentration (SPM) was the only variable related to systematic variations in partitioning of trace metal and Fe and Al concentrations between filter-retained and filter-passing fractions across salinity gradients. Inverse relationships were observed between empirically defined particle/solution distribution coefficients, KD, for the different metals and SPM concentrations. This inverse dependence can be explained by the “particle concentration effect”, which can be caused by the presence of Fe, Al, and trace metals associated with colloidal matter in the filtrate fraction. Our argument is supported by the direct analysis of colloids (> 10,000 Daltons) ultrafiltered from Galveston Bay waters. Colloidal Fe, Al, Pb, and Zn (but not Cu) account for most of the filter-passsing form of these metals. The results of this study have important implications for regulatory agencies and regulated industries, allowing for algorithms to predict Pb, Cu, Zn, and Ag partitioning between particle and solution phases.

The influence of size distribution on the particle concentration effect and trace metal partitioning in rivers

The particle concentration effect (p.c.e.) is an unexpected decline in partition coefficients (Kd) as suspended particulate matter (SPM) increases. This anomaly has been attributed to a variety of causes, but most often to the existence of colloidal forms of the adsorbate, which are included, in error, in the dissolved fraction when calculating Kd. To test this hypothesis we have directly measured colloidal, macroparticulate, and truly dissolved metals (Cd, Cu, Pb, Zn, Fe, Al, Mn, Ag) monthly for one year (6/96–7/97) in four Connecticut rivers having a range of ancillary biogeochemical characteristics. These include factors that are all likely to influence partitioning between dissolved and solid phases, such as DOC, pH, and competing cations. The p.c.e. is clearly evident in these rivers, and explicit consideration of colloidal metals eliminates or reduces this anomaly in nearly all cases. Furthermore, a substantial portion of metals occurs in the colloidal fraction, and the amount of colloidal metals increases with SPM. Both of these conditions are necessary for the colloidal model to explain the p.c.e.. Important differences are observed among rivers and metals, and in some cases systematic decreases in Kd continue even when colloidal metals are taken into account. This additional decrease can be eliminated by excluding large particles having negligible surface complexation sites. When corrections are applied for both colloids and large particles, Kd values become truly constant.

Scavenging of thorium isotopes by colloids in seawater of the Gulf of Mexico

A suite of surface-water samples from the Gulf of Mexico was analyzed to ascertain the extent of association of Th isotopes (232Th, 234Th) with colloids and the role of colloidal material in scavenging Th from the water column. These are the first measurements of naturally occurring colloidal Th. The fraction of 232Th, 234Th associated with colloids (i.e., 10,000 Dalton < colloids < 0.4 μm) ranged from 10 to 78% of the Th passing 0.4 μm Nucleopore cartridge filters. Colloid mass concentrations were much larger than the corresponding 0.4 μm filter-retained particle concentrations. The conditional partitioning constants for 234Th with colloids, Kc, and filter-retained particles, Kf, are comparable in magnitude. Thorium scavenging rate constants decreased in value with increasing distance from the coast (0.164 to 0.007 d−1), and this is attributed to the decreasing particulate-matter concentration from coastal to deeper Gulf waters. In addition, there exists a highly significant, positive correlation between values of the Th scavenging rate constant and the fraction of 0.4 μm filter-passing Th associated with colloids. An average residence time of 6 days, with a range of 4–26 days, was calculated for the characteristic time scale of colloid transfer through the particle size spectrum, including sedimentation. The large fraction of 234Th which was found to be associated with colloids suggests that Th isotopes can be used as in-situ “coagulometers,” tracing the aggregation of colloidal material with, or into, large aggregates of filter-retained sizes.

Rapid, plant-induced weathering in an aggrading experimental ecosystem

To evaluate whether rates of weathering of primary minerals are underestimated in watershed mass-balance studies that fail to include products of weathering accumulating in plants and in developing soil, changes in the calcium and magnesium content of vegetation and soil fractions were measured in large, monitored lysimeters (sandbox ecosystems) at Hubbard Brook Experimental Forest, New Hampshire. Weathering was evaluated over 4–8 yr in sandboxes planted with red pine (Pinus resinosa Ait.) and kept mostly free of vegetation (nonvegetated). Three mass-balance equations were used that cumulatively include (a) Ca and Mg in precipitation inputs and drainage outputs, (b) accumulation of Ca and Mg in vegetation, and (c) changes in products of weathering in soils. Soil products were evaluated with an extraction process designed to avoid removing ions from primary minerals. Relative to the input-output equation, the estimated rate of weathering increased 2.4 (Ca) and 1.8 (Mg) times when accumulation of Ca and Mg in pine biomass was accounted for, and 8 (Ca) and 23 (Mg) times when changes in soil products were also included. Weathering estimates that included accumulation in vegetation and soil products were 261 (Ca) and 92 (Mg) kg ha-1 yr-1 in the pine sandbox. These rates were 10 (Ca) and 18 (Mg) times higher than the rates in the nonvegetated sandbox, which were not significantly greater than zero. This study raises the possibility that weathering can play a significant role in the release of nutrients available to plants over short periods. Faster rates like this become extremely important where managers are trying to balance nutrients available to plants from precipitation and weathering release with outputs including harvest removals.

Evidence of the particle concentration effect for lead and other metals in fresh waters based on ultraclean technique analyses

Ultraclean methods were used to produce reliable concentration data for the trace metals Pb, Ag, and Cd in fresh waters and for Ph, Ag, Cu, and Zn in estuarine waters. Partitioning of metals between filter-retained and filtrate fractions exhibited a dependence on total suspended solids (TSS) concentration. This phenomenon, the particle concentration effect (PCE), has been previously documented almost exclusively in marine and estuarine systems and lab simulations, and mainly for radionuclides. The partition coefficient, Kd, was independent of major ion chemistry and pH, supporting the hypothesis that the PCE is caused by metals associated with colloidal particles but counted with the filtrate (“dissolved”) fraction. Partition coefficients of the measured metals in fresh waters are predictable across the full range of TSS measured, spanning more than two orders of magnitude. The inferred true partition coefficient for Pb (between solution and particles of all size classes) is greater than 107.4, suggesting that truly dissolved Pb concentrations are extraordinarily low. Previously published data are reinterpreted to show that naturally occurring 210Pb also exhibits the PCE. Freshly precipitated Fe oxyhydroxides partition metals exactly like organic detritus and clays in spite of the great difference in their surface chemistry. The same data rule out the possibility that the PCE could be caused by a decrease in surface area (and surface complexation sites) due to resuspension of larger particles under high TSS conditions. A surprising result is that, while the slope of a log (Kd) — log (TSS) plot for 210Pb is the same as for stable lead, absolute Kd values for 210Pb are uniformly lower by a factor of 4. This suggests that 210Pb and stable lead behave differently from each other in the surface waters studied. One possible explanation is that this dissimilarity may be attributable to differences in speciation that are persistent on a time scale of months, corresponding to the water residence time or Pb removal rate.

Clean technique measurement of pb, ag, and cd in freshwater: a redefinition of metal pollution.

Ultraclean sampling, handling, and analysis techniques are necessary to obtain reliable trace metal data for aquatic environments. Failure to follow appropriate clean protocols calls into question much of the previous research on metal cycling in freshwaters (1-41, just as virtually all marine trace metal data from before about 1975 are now considered invalid. Reported here are data for Pb and Cd in freshwaters measured using clean techniques and some of the first such data for Ag. Samples were collected from the length of the Quinnipiac River, Connecticut, and its tributary streams. At all locations, Pb, Ag, and Cd were below the detection limits of routine monitoring measurements by governmental agencies. In spite of these lower than expected levels, the metals are almost 2 orders of magnitude higher in the industrialized portions of the river than in the undeveloped headwater streams in this watershed or elsewhere in New England, and clear reproducible trends are evident. Only through the use of clean techniques and sample preconcentration is it possible to detect the enormous difference in metal concentration levels between clean and contaminated portions of the river. Likewise, only if these methods are strictly applied will it be possible to monitor trends in toxic trace metals over time. This study and other recent work suggest that we may need to redefine the level a t which a river is considered “polluted” with heavy metals.

Sedimentation rates in flow-restricted and restored salt marshes in Long Island Sound

Many salt marshes in densely populated areas have been subjected to a reduction in tidal flow. In order to assess the impact of tidal flow restriction on marsh sedimentation processes, sediment cores were collected from flow-restricted restricted salt marshes along the Connecticut coast of Long Island Sound. Cores were also collected from unrestricted reference marshes and from a marsh that had been previously restricted but was restored to fuller tidal flushing in the 1970s. High bulk densities and low C and N concentrations were found at depth in the restricted marsh cores, which we attribute to a period of organic matter oxidation, sediment compaction, and marsh surface subsidence upon installation of flow restrictions (between 100 and 200 years before the present, depending on the marsh). Recent sedimentation rates at the restricted marshes (as determined by137Cs and210Pb dating) were positive and averaged 78% (137Cs) and 50% (210Pb) of reference marsh sedimentation rates. The accumulation of inorganic sediment was similar at the restricted and reference marshes, perhaps because of the seasonal operation of the tide gates, while organic sediment accretion (and pore space) was significantly lower in the restricted marshes, perhaps because of higher decomposition rates. Sedimentation rates at the restored marsh were significantly higher than at the reference marshes. This marsh has responded to the higher water levels resulting from restoration by a rapid increase in marsh surface elevation.

Radiocarbon dating of a core from Long Island sound

Abstract Radiocarbon measurements sequentially with depth in a diver-obtained core from Long Island Sound has been interpreted in terms of sediment accumulation rates and sources of organic carbon in the sediment. A sediment accumulation rate of 0·075 ± 0·013 cm year −1 is determined from the data below 10 cm in the core. An age of 2320 year BP is obtained for the sediment-water interface by extrapolation. This means that the dominant carbon component preserved in the core is soil-derived. The contributions of fossil fuel carbon and bomb radiocarbon associated with plankton are also evaluated.

Evidence for diffusive redistribution of 210Pb in lake sediments

Abstract 210 Pb, 210 Po, and ancillary geochemical parameters were measured in the sediments and pore waters of a lake with seasonally anoxic bottom waters. Substantial release of radionuclides to the water column has been documented at this site. Solid phase 210 Pb profiles do not match the expected input history, suggesting that the radionuclide may be undergoing redistribution. High levels of the radionuclides were measured in pore waters, consistent with partition coefficients in the range from 10 3 to 10 4 . The high pore water activities, apparent redistribution pattern, and the documented release of 210 Pb from these sediments to the water column, all point to the possible importance of pore water diffusion as a 210 Pb transport mechanism. The distribution of 210 Pb in these sediments was successfully modeled using a combination of sediment burial and pore water diffusion without the need to invoke particle reworking. Theoretical analysis supports the idea that in some cases large dating errors can result if diffusive redistribution of 210 Pb is neglected.