A 2100-year trace-element and stable-isotope record at decadal resolution from Rice Lake in the Northern Great Plains, USA

A 2100-year multiple proxy record at decadal resolution from a topographically closed (but hydrologically open) lake basin was used to reconstruct changes in lake salinity and evaporative intensity and to evaluate the reliability of different proxies for inferring climatic change in the glaciated Northern Great Plains (NGP). At Rice Lake, North Dakota, the ostracode-Mg/Ca ratios show good correlation with other ostracode, diatom and aeolian proxy records from other sites in the NGP. All of these records show significant century-scale periodicities, which have been related to solar variability as inferred from atmospheric radiocarbon records. This coherent and consistent pattern suggests that these sites record climatic variability at regional scales. The interval of the Mediaeval Climatic Anomaly from ∼900 to 550 cal. yr BP contains two pronounced dry periods. During the following ‘Little Ice Age’, a single period of frequent drought that peaked at ∼300 cal. yr BP was bracketed by wet periods before and after. In contrast, the ostracode-Sr/Ca ratios and ostracode- and mollusc- δ¹⁸O records from Rice Lake have different patterns and do not correlate well with each other or with the same proxies at other sites. This lack of consistent pattern within a site and among several regional sites suggests seasonal and/or site-specific influences on these proxies. Variable contributions of groundwater, with low δ¹⁸O values because of selective recharge from snowmelt, may have played an important role in determining the isotopic composition of the lake. The poor correlation of Sr/Ca with other proxies is caused by mediation of inorganic carbonate mineralogy, owing to active uptake of Sr in the formation of inorganic aragonite. Thus, our sediment data along with present-day lakewater chemistry suggest that carbonate-δ¹⁸O records in the NGP cannot be interpreted simply in terms of climate because of significant local groundwater influences. Ostracode-Mg/Ca ratios tend to be a better indicator of past salinity because of the conservative nature of Mg in oligosaline waters (<10‰ salinity).