VentDB Help: More Information
Please click on a topic for more information:
- Units for Concentration of Chemical Species
- Zero-Mg Hydrothermal End-Member Compositions
- Quality Control Parameters for Zero-Mg End-Member Composition Estimates
- Lowest Mg concentration, Mg(min)
- Number of samples (n) and correlation coefficient (r^2)
- Charge Balance
- Searchable List of Chemical Elements and Species
- Types of Sampling Device
- Latitude, Longitude, and Depth
In comparing chemical data from one site or study to another, it is essential that concentrations be expressed in the same units. The preferred unit, and that most used by far in the scientific literature on seafloor hydrothermal vents, is moles per kilogram of solution. This is in spite of the fact that many analyses of vent water, especially those performed at sea, are performed volumetrically, in moles per liter of solution, because of the difficulty of using an analytical balance on a rolling ship. The reason for this discrepancy is that, unlike for unaltered seawater, hydrothermal solutions can have widely varying density, depending on their pressure and temperature, and in modeling the physical and chemical behavior of such solutions in nature it is mass which is conserved and not volume. In VentDB we follow the scientific community in using moles per kilogram of solution, the usual unit in the publications we have accessed. In those rare instances where concentrations are given in moles per liter of solution, we convert to moles per kilogram by dividing the volumetric concentrations by the somewhat arbitrary value of 1.0243, approximately the density of seawater at 22°C, a common laboratory temperature. As many of the published analyses have been performed volumetrically, yet are presented per unit mass, the analysts or authors have presumably performed some similar calculation in the opposite direction, although this is almost never acknowledged, let alone detailed.
Because submarine hot springs vent into the overlying ocean, it is difficult to sample the venting hot water without some contamination from ambient seawater. Samples collected are thus typically mixtures of the hydrothermal end-member and bottom seawater. Admixture can occur above the seafloor during sampling, within a chimney through leaks at its base and along its sides, and even within the shallow seafloor itself, as was first documented by Edmond et al. (1979a,b) at the original discovery site, the warm springs on the Galapagos Rift near 86°W. Provided that mixing has occurred shortly before sampling, the seawater end-member will be unreacted for all but the most reactive chemical species, and most species from both seawater and vent water will tend to behave conservatively on mixing. This contention has been confirmed repeatedly, at many vent sites on the seafloor, by the observation of linear relationships between one chemical species and another in multiple samples collected from a single vent or vent field. Edmond et al. (1979a) first demonstrated these linear relationships, which they interpreted as mixing lines, by plotting various chemical species against dissolved silica. Since then it has become conventional to plot various species against the concentration of Mg, because Mg is present at high concentration (~52 mmol/kg) in seawater, and because laboratory experiments reacting seawater with basalt have demonstrated that at temperatures of ~200°C and above, Mg is almost completely removed from seawater into alteration minerals, such that the remaining concentration is <1mmol/kg (Bischoff and Dickson, 1975; Mottl et al., 1978; Seyfried and Bischoff, 1978?). For this reason it has become conventional to report the hydrothermal end-member composition for a vent or vent field as that determined by extrapolating from the ambient seawater composition, through the compositions of the individual samples collected, to a Mg concentration of zero.
The end-member hydrothermal compositions reported in VentDB are those based on extrapolation to zero-Mg. In nearly all cases these extrapolations were made by the authors of the papers from which the data have been taken. These zero-Mg compositions are usually the only values that are tabulated in published papers, although the raw data from individual samples are often shown in plots of various chemical species against Mg. Where possible we have obtained the raw data, usually from the original authors themselves, as well as the end-member compositions, and have tabulated both in VentDB. In rare cases in which the original authors have published raw data and not zero-Mg end-member compositions, we have made these extrapolations ourselves and included the results in VentDB. The actual extrapolations can be found in the Excel spreadsheets produced by M.J. Mottl of the University of Hawaii which were used as data input to VentDB. These spreadsheets reside in the Geochemical Reference Library and can be searched by “Mottl”. In the future we plan to provide links to them in VentDB.
Because some vents and vent fields have been revisited and sampled in different months and years, we have appended the year to the name of the vent or vent field in designating the hydrothermal end-member composition, as this information is necessary in assessing whether the composition has changed with time.
Besides the usual quality control parameters of precision and accuracy that accompany the chemical analyses, we have included a set of four parameters that apply to the end-member hydrothermal compositions, estimated by extrapolation to a Mg concentration of zero. These parameters are typically not included in the original publications but represent value-added feature of VentDB. They are 1) the lowest Mg concentration (Mgmin) measured in a set of samples from a vent or vent field; 2) the number of samples (n) used in the extrapolation; 3) the correlation coefficient (r2), usually for the fit of Ca vs. Mg; and 4) the electrical charge balance for the extrapolated, hydrothermal end-member composition.
Submarine hot springs with temperatures above ~200°C can be expected to have a Mg concentration of <1 mmol/kg. Samples with higher Mg concentrations can almost always be interpreted as containing substantial admixture of ambient seawater, entrained during or shortly before sampling, above or within the shallow seafloor or venting chimney. (See Ravizza (20??) for a purported exception.) Within a set of samples from a vent or vent field, the one with the lowest Mg concentration is thus the least contaminated by seawater, and represents the purest sample of hydrothermal end-member water collected from the vent. The conventional extrapolation to zero-Mg is quite short, and often insignificant, for such samples. In addition, low-Mg samples are the only ones that can provide reliable data for concentrations of species that highly reactive on mixing, such as many transition metals, which typically produce non-linear relationships when plotted against Mg. Low-Mg samples are thus highly prized, and end-member hydrothermal compositions based on very low-Mg samples are more reliable than those for which the lowest Mg concentration is higher. The lowest Mg concentration measured in a sample suite is thus the most important indicator of accuracy for an estimated hydrothermal end-member composition.
For a set of water samples collected from a vent or vent field, the best-fit line through ambient seawater and the concentration of a given chemical species measured in the samples, when plotted against Mg, yields an assessment of a) the spatial and temporal uniformity of the hydrothermal end-member, b) the reactivity of the chemical species on mixing, and 3) the accuracy of the estimated composition of the hydrothermal end-member. A larger number of samples (n) and a higher correlation coefficient (r2) imply a more accurate estimate. Regression lines can be fitted for every chemical species against Mg. Where possible, we provide these values for Ca vs. Mg, because Ca is a major element in seawater that is virtually always measured and that usually behaves conservatively on mixing. The actual plots can be found in the Excel spreadsheets produced by M.J. Mottl of the University of Hawaii which were used as data input to VentDB. These spreadsheets reside in the Geochemical Reference Library and can be searched by “Mottl”. In the future we plan to provide links to them in VentDB.
Where possible, we provide the electrical charge balance, calculated as the sum of cations vs. anions, for every composition, whether based on raw data or end-member extrapolation to zero Mg. For the raw data, derived from analysis of individual water samples, authors have often calculated the Na concentration by difference using charge balance, and report this calculated concentration rather than a measured one. This is a common practice because Na and Cl are by far the major ions in seawater, and the most common method for determining Cl (electrochemical titration with silver nitrate, which determines chlorinity, which is chloride plus bromide: ±0.4%, 1-σ) is much more precise than the most common method for determining Na (ion chromagraphy: ±2%). Where Na has been calculated from charge balance, the charge balance we calculate should be zero. It rarely is, though, possibly because of rounding errors, so we report the charge balance we calculate from the published data in any case. For the hydrothermal end-member compositions, charge balance is a useful check on the quality of the extrapolated composition, because, regardless of how Na has been determined, the concentration of each species in the hydrothermal end-member has been estimated by a separate extrapolation against Mg. If significant errors in the concentration of charged species have been introduced by these extrapolations, they often reveal themselves by an imbalance in electrical charge.
The list of chemical elements and species that can be searched for in VentDB includes headings for Key Properties (temperature and quality control parameters), Chemical Species, Dissolved Gas, Radiogenic Isotopes, and Stable Isotopes. Inclusion of a given species in one list or the other is somewhat arbitrary: e.g., are H2S and NH3 chemical species or dissolved gases? (We put one in each.) The order in which chemical species are listed is also somewhat arbitrary and represents a compromise between abundance in vent waters and order in the Periodic Table of the Elements. We lead with pH, alkalinity, the six major ions in seawater, and Si. Chlorinity (equals chloride plus bromide), as measured by titration with silver nitrate, and chloride, as measured by ion chromatography, are listed separately. We follow with the rest of the halides, alkalies, and alkaline earth elements, and then B and the rest of Column 3 along with Ge from Column 4. Next come the abundant hydrothermal metals Mn, Fe, Cu, and Zn, followed by eleven trace metals that can be important in hydrothermal solutions and are sometimes determined, and 15 generally less abundant transition metals that have seldom if ever been determined. Following them are Column 5, including N and P species, four trace elements from Columns 6 and 7, the 15 lanthanide elements, and the four actinides. Along with H2 and O2 and the noble gases in the Dissolved Gas list, this accounts for 91 chemical elements and numerous species; only Tc is not included, as it does not occur in nature.
Several types of sampling devices have been used to collect hydrothermal solutions on the seafloor, which vary by their material, design, volume, and whether they retain dissolved gases or not. Some types can be used either separately or as part of a flow-through manifold mounted on a submersible, which allows temperature and other parameters to be measured in the flow stream prior to and during filling of the sampler. Some samplers were used only once, in the early days of vent research in the 1970’s. The most frequently used has been the Walden-Weiss Ti-syringe sampler, which has the disadvantage that dissolved gases are not retained during the pressure decrease that accompanies recovery from the deep sea to the sea surface. A short description of several types of sampling devices follows.
- Acrylic pumped sampling manifold: used on the Galapagos Rift in 1977. (Corliss et al., 1979, Science 203, 1073-1083)
9-L clear acrylic cylinders, 6 each mounted on a pumped manifold. Warm spring water is pumped through the manifold and past a large (~20 cm) diameter filter and into the plexiglass cylinder. Downstream the spring water flows past sensors for temperature, conductivity, dissolved oxygen, and pH.
This device was designed and built by Milo Clauson at Oregon State University. It was used from 1977-1979 on the Galapagos Rift near 86°W, from the manned submersible Alvin. In 1979 the downstream sensors for temperature, conductivity, dissolved oxygen, and pH were not installed, as noted by Lilley et al. (1983, p. 413).
- Los Alamos sampler, designed by Jake Archuleta: used on the EPR 21°N in 1979. (Von Damm et al., 1985, Geochim. Cosmochim. Acta 49:2197-2220) Built for the Los Alamos Hot Dry Rick Geothermal Power Program. 750-mL stainless steel bottles, internally gold plated, evacuated, with ¼-inch inlet tube. Too heavy to deploy with Alvin’s arm, so that the whole sub had to be maneuvered. Collected 14 samples in 1979, of which only two contained >50% hydrothermal solution (Von Damm et al., 1983, in Rona volume; Edmond et al., 1982, Nature 297:187-191).
- Walden-Weiss titanium-syringe sampler, first used on the EPR 21°N in 1981. (Von Damm et al., 1985, Geochim. Cosmochim. Acta 49:2197-2220) Designed by Barrie Walden of WHOI and Ray Weiss of SIO. A pair of 755-mL titanium syringes mounted in parallel on a T-handle, each filled through its own 0.5-inch titanium snorkel. The syringe walls are of thin titanium and the piston head is relatively massive, maximizing heat loss and allowing the use of Teflon omni-ring seals. Spring loaded; requires about one minute to fill. Prevented precipitation in 1981 except when >15% seawater was entrained. Dead volume is 3.8 mL (0.05%), which was filled with surface seawater for the 1981 dives. Ten hours typically elapsed between collection and processing. Once started, the solution may be removed from the sampler in 15 minutes or less. One syringe was used for chemistry and the other for stable isotopes and dissolved gases. The chemistry aliquot was acidified immediately with 6N HCl to a pH of about 1.6.
Samples were filtered onshore through 0.4 um Nucleopore filters. The filtrate was digested in 8N HNO3 and liquid Br2 and also analyzed. Larger particles of chimney and smoke entrained during sampling apparently settled to the bottom of the syringe and thus were not redissolved after filtration.
- NOAA manifold sampler, used with Walden-Weiss titanium-syringe samplers, first (?) used on the JFR in 1987. (Massoth, G.J., H.B. Milburn, S.R. Hammond, D.A. Butterfield, R.E. McDuff, and J.E. Lupton, 1989, The geochemistry of submarine venting fluids at Axial Volcano, Juan de Fuca Ridge: new sampling methods and a VENTS program rationale. Natl. Undersea Research Program Res. Rept. 88-4, pp. 29-59, NOAA, Wash. D.C.) Designed and built at NOAA/PMEL by Gary Massoth and Hugh Milburn. A pumped, flow-through manifold sampling device which measures temperature on-line during sampling, both upstream and downstream from the sampling bottle, to minimize entrainment of ambient seawater. Used in conjunction with a sample bottle, often the Walden-Weiss Ti-syringe samplers. Only when temperature is stable at approximately the maximum vent temperature are the titanium-syringe samplers triggered to collect the sample.
- Lupton gas-tight sampler, designed by John Lupton of NOAA/PMEL in Newport, Oregon. Seewald Isobaric Gas-Tight (IGT) sampler, designed by Jeffrey Seewald of Woods Hole Oceanographic Institution.
These essential metadata for samples, vents, and vent fields are often omitted in publications that present hydrothermal vent compositions. Where not provided, we have collected this information from other sources whenever possible. Our preferred source, especially for hydrothermal end-member compositions for various vents and vent fields, is the Marine Geoscience Data System (MGDS), which compiles location data from a variety of sources and often presents carefully renavigated locations determined over multi-year visits to various sites. For individual samples collected by manned or unmanned submersible we have often used the DSV Alvin and ROV Jason dive logs maintained online by the Woods Hole Oceanographic Institution. For older samples a single location for a dive is usually all that is readily available, along with a depth that is the deepest reached on a given dive. When no more accurate depth is available for a vent or vent field we use this depth, recognizing that the true depth may be shallower. When a vent or vent field has been visited on several dives, we use the shallowest of the maximum depths for these several dives, as provided in the online dive logs.