PALEOSALINITY FROM TRACE-METALS IN FOSSIL OSTRACODES COMPARED WITH OBSERVATIONAL RECORDS AT DEVILS LAKE, NORTH-DAKOTA, USA

The chemistry of closed-basin lakes responds directly to the hydrologic budget through evaporative concentration of dissolved salts, and the sediments of these basins contain several excellent fossil and geochemical proxies for past changes in water chemistry and salinity. The trace-metal content of ostracode shells, among the most promising of these paleosalinity methods, is further explored in this study of Mg and Sr partitioning in Candona rawsoni, a widespread ostracode of the North American Great Plains. The molar distribution coefficients K(D)[Mg] and K(D)[Sr] for this species, [GRAPHICS] [GRAPHICS] are determined from laboratory experiments in which C. rawsoni are cultured under controlled temperatures and salinities. Our formulation of K(D)[Mg] contains a correction term for excess Mg that is incorporated into the shell during early calcification. The accuracy of the distribution coefficients is tested against field collections of live ostracodes from a suite of 12 Dakota lakes. The K(D)-inferred values for (Mg2+/Ca2+)H2O and (Sr2+/Ca2+)H2O are in close accord with the same ratios measured in the lakes. A comparison of ostracode-reconstructed and historically measured salinities for Devils Lake, North Dakota, shows good agreement within the salinity range preferred by C. rawsoni (1-10 parts per thousand), although salinity excursions beyond this optimum are poorly represented in the sedimentary record. In most lakes salinity is more readily inferred from (Mg/Ca)H2O than (Sr/Ca)H2O because Sr solubility is a complex function of several carbonate phases. As shown for Devils Lake, (Sr/Ca)H2O may not vary systematically with the salinity of some brines.