reaction being monitored, and the importance of model assumptions to the flux calculation.
Most benthic flux estimates made to date have been obtained by applying transport models to concentration versus depth profiles of pore water solutes. An important feature of these estimates is that they are only as accurate as the transport parameters needed to convert the concentration measurements into flux estimates. In pelagic sediments, the dominant transport process is molecular diffusion; here, accurate estimates of sediment diffusivity and porosity near the sediment-water interface are required. In continental margin sediments, solute transport is enhanced over that due to diffusion by biologically and physically driven irrigation, and models applied to pore water profiles must include independent estimates of the effects of these processes (Aller, 1980b; Christensen et al., 1984). Benthic flux estimates that are less dependent on transport models are obtained through the use of benthic flux chambers. In these experiments, sediments and a small volume of overlying water are enclosed for incubation periods of hours to about a month, and changes in concentration of diagenetic products or reactants in the overlying water are monitored. It appears that, with the possible exception of reactions with scale lengths on the order of a millimeter, the assumptions leading from these measurements to benthic flux estimates are reasonably robust (e.g., Bender et al., 1989). In the special case of rapid reactions occurring primarily within one or two millimeters of the sediment-water interface, neglect of the diffusive sublayer at the sediment-water interface may lead to erroneous flux estimates (Boudreau and Guinasso, 1982). Benthic flux estimates based on flux chamber experiments are probably preferable to pore water profile based estimates because, in most cases, the estimates are essentially model-independent.
A second consideration that is important to evaluating benthic flux estimates is the resolution of the method used. For pore water profile based estimates, accurate flux calculations depend on the relationship between the length scale over which the diagenetic reaction occurs in the sediments and the sampling length scale. Because many reactions occur on length scales of millimeters to one or two centimeters in many locations (Emerson and Bender, 1981; Bender and Heggie, 1984; Reimers, 1987; Bender et al., 1989), sampling resolution is an important limitation of many pore water based flux estimates. "Traditional" core sectioning and centrifuging/squeezing methods have been used with a resolution of 0.5 to 2 cm. In situ sampling methods based on "harpoon"-type samplers (Sayles et al., 1973b), while free of the artifacts associated with core recovery, have been limited to a sampling resolution of 2 to 5 cm. Recently, two methods capable of in situ measurement of several pore water components have begun to be used: electrode techniques for the measurement of dissolved oxygen and pH (Revsbech et al., 1980; Reimers
et al., 1986; Archer et al., 1989), and whole-core squeezers for measuring solutes whose concentration is not changed by disturbances of ion exchange equilibria on very short time scales (minutes) (Bender et al., 1987; Martin et al., 1988; Bender et al., 1989). Both of these methods are capable of millimeter-scale resolution. We compare the resolution of several of these techniques in Figure 10.1. For flux chamber based estimates, one must consider whether the experiment length is sufficient to produce a measurable concentration change in the solution component of