CADMIUM VERSUS PHOSPHATE IN THE WORLD OCEAN

Cadmium (Cd) is one of the best studied trace metals in seawater and a t individual stations exhibits a more or less linear relation with pho sphate. The compilation of all data from all oceans taken from over 30 different published sources into one global dataset yields only a bro ad scatterplot of Cd versus phosphate. However, the smaller high-quali ty dataset obtained by rigorous selection of only those stations with uniform Cd/PO4-ratio in the deep waters, provides a consistent global description of the deep (> 1000 m) waters. The deep Cd/PO4-ratio incre ases from about 0.18 X 10(-3) in the subarctic North Atlantic to about 0.33-0.35 x 10(-3) in the northern Indian and Pacific Oceans, in acco rdance with increasing phosphate content, i.e. age, of the deep water. The increasing Cd/PO4-ratio with age (and phosphate) of the deep wate r masses is a function of the coupling between biogeochemical cycling and deep water circulation. Changes in the latter, for example during a glacial period, inevitably lead to significant shifts in the Cd/PO4 relationship of seawater. There is a statistically significant bimodal ity of deep Atlantic versus deep Antarctic/Indo/Pacific waters, sugges ting that the deep Atlantic is a distinct biogeochemical province for Cd cycling. This distinction is likely caused by the high inventories of both Cd and phosphate in Weddell Sea source waters. For each of bot h populations, a given concentration of phosphate yields a predicted v alue of Cd within +/- 100 pM (Atlantic) and +/- 200 pM (Antarctic/Indo /Pacific), respectively, at the 95% confidence level. If one ignores t he bimodality, then for a given phosphate the corresponding Cd might b e predicted within +/- 150 pM at the 95% confidence level; the validit y of this is currently being verified by studies of South Atlantic wat ers which may or may not provide the missing link between both populat ions. Currently, the global distribution of the Cd/PO4-ratio in surfac e, thermocline and deep waters is consistent with preferential biogeoc hemical removal of Cd versus phosphate from surface waters. The net re sult for Cd/PO4 is not dissimilar to the preferential surface removal of C-12 over C-13 driving the deep distribution of the dissolved C-12/ C-13-ratio, although for Cd/PO4 the underlying mechanism is obviously very different and not well understood.