A scavenging model for trace elements in the deep sea

A vertical diffusion-advection model with a scavenging removal rate which is first-order in concentration is applied to the deep-water Geosecs-I profiles of four trace metals. The actual model scavenging rates are obtained by calibration of the vertical advection velocity with the radiocarbon profile from the same station — the¹⁴C data givew = 3.7m/yr. “Scavenging residence times”, τ_ψ, of 1400 years for Cu, 3200 years for Sb, and 2500 years for Sc are obtained. For these three elements the fit of the data to the scavenging-model equation is better than the fit to a conservative mixing model with no in-situ removal. TheF_χ comparison test is used to assess the probability that the improved fit given by the scavenging model is due merely to a random selection of data from an infinite set of values with a distribution governed by the stable-conservative model. This probability is found to be only 0.6% for Cu and only 3% for Sb. For Sc, however, the probability is 19% so that the two models cannot be distinguished for this element. The Ni profile shows no effect of scavenging; a lower limit of 3000 years is estimated for τ_ψ but the value is probably much greater. The association of Cu and Pb as two elements with definite indication of scavenging in deep water is consistent with experimental adsorption studies; the lack of indication of scavenging of Ni is consistent with the hypothesis that clay minerals are the scavenging agent for Pb and Cu. The method of comparing scavenging and conservative mixing-model fits to the profiles provides a mechanism for defining a set of trace elements which show consistent scavenging effects over large ocean areas, so that the reality of the deep-sea scavenging process for non-radioactive elements can be tested.