DETERMINATION OF THE OPTIMUM SAMPLING INTERVALS IN SEDIMENT PORE WATERS USING THE AUTOCOVARIANCE FUNCTION

Abstract

An approach to scaling that has found considerable utility in the field of physical oceanography is based on the auto-covariance function (ACF). It is demonstrated to also be useful in determining the characteristic length scale of the dominant variance of dissolved sulfide and ferrous iron in anoxic pore waters. The method begins with the analysis of a data series in which the measurement sampling interval is smaller than the expected length scales. A least squares polynomial fit to the data representing a background field is then removed from the data series making possible the analysis of the variability superimposed on the background field. The characteristic length scale is then defined as the distance in which the measurements become uncorrelated. This usually occurs at a significance-level or zero-crossing of the ACF. Characteristic lengths obtained for sediments from a variety of coastal environments are not distinguishable between second and third order polynomial fits with average values and standard deviations of, respectively, 7.0 ± 2.7 mm and 6.1 ± 2.3 mm. These values are close to the typical burrow diameter of 7.5 mm used by Aller (1978) in his bioirrigation model for sediment diagenesis. These results indicate that macrofauna dwelling in the sediment are probably responsible for the variability, although in the seagrass meadow rhizomes may also play a role.