interest. Ultimately, experiment length is limited by the need to maintain incubation conditions close to those of the steady state sediment system. In most sediments, experiment length is limited by changes in the bottom water oxygen concentration or by the availability of labile organic matter in the closed system; these limitations make it difficult to measure fluxes of slowly reacting components of the system, particularly if they are present at high concentrations in bottom water. Thus, fluxes of major ions, and, to a lesser extent, ΣCO2, are difficult to measure using benthic flux chambers.
Because the changes in concentration that must be resolved to estimate benthic fluxes are often small, samples must be artifact-free. Artifacts introduced in the recovery of cores from the seafloor are known to significantly alter pore water concentrations for many of the major species of interest here. These artifacts occur as a result of warming during recovery and can strongly affect the major cations through ion exchange reactions (Mangelsdorf et al., 1969; Bischoff et al., 1970; Sayles et al., 1973a). De-pressurization during recovery of cores from the deep sea can also introduce artifacts through CaCO3 precipitation, substantially changing ΣCO2, alkalinity, and Ca2+ concentrations (Murray et al., 1980; Emerson et al., 1982). It is these problems that led to the development of in situ pore water sampling techniques (Barnes, 1973; Sayles et al., 1973b). Comparisons of profiles obtained by in situ and shipboard techniques have shown that shipboard techniques can be used for measuring dissolved oxygen, nitrate, and silicate; they do not appear to be suitable for phosphate measurement (Jahnke et al., 1982b). Because benthic flux chamber measurements are typically carried out in situ, they are free of artifacts due to shipboard separation of pore waters from the solid phases of sediments.
Sampling resolution, measurement artifacts, and transport model assumptions affect fluxes calculated from different measurement techniques to varying extents. There is another important class of model assumptions which affect all flux calculations equally. These assumptions result from our limited knowledge of diagenetically important chemical reactions. The outstanding example, one which will be important to our subsequent discussion, is the calculation of the rate of organic carbon degradation at the sea floor. Because of the difficulty of making artifact-free measurements of dissolved inorganic (and organic) carbon in pore waters, benthic organic carbon degradation rates are often estimated based on fluxes of nitrate and oxygen across the deep sea sediment-water interface, and on fluxes of sulfate in nearshore sediments. Translations of these measurements into organic carbon fluxes require assumptions about the oxidation/reduction stoichiometry of organic carbon degradation reactions (see Table 10.1). The oxidation states of C, N, and H in organic matter are
TABLE 10.1 Organic Matter Degradation Reactions
generally represented by assuming the elements are present as CH 2O and NH3. All of the the organic C is assumed to be oxidized to CO2; depending on the environmental redox potential, N is oxidized to NO3- (oxic degradation) or N2 (denitrification and Mn reduction), or released as NH3 (Fe reduction, sulfate reduction) (Froelich et al., 1979). Takahashi et al. (1985) have called into question the assumptions about the oxidation states of the reactants, par-