Trace-element analysis of the calcareous shells of ostracodes in a sediment core from Farewell Lake provides the first limnogeochemical record for climatic reconstructions in Alaska. When compared with pollen data from the same site, this record offers new insights into climatic controls over vegetation dynamics during the Holocene. The low Mg/Ca ratios and high Sr/Ca ratios suggest that a relatively cold dry climate prevailed in this region between 11,000 and 9000 yr B.P. (uncalibrated 14C ages are used throughout the paper). This result contrasts with previous interpretations of a thermal maximum at this time, corresponding to the widespread establishment of Populus woodland/forest. The trace-element record suggests, instead, that the warmest period of the early Holocene at Farewell Lake was between 8500 and 8000 yr B.P. during the decline of Populus. Marked decreases in Sr/Ca and Mg/Ca suggest a major increase in effective moisture around 6500 yr B.P., which coincided with the establishment of Picea boreal forests in the Farewell Lake region. This climatic change was probably widespread throughout much of Alaska and adjacent Canada and might have induced the rapid spread of Alnus and the shift from Picea glauca to P. mariana dominance across that region. Our geochemical record also suggests that the late-Holocene climate history was more complex than previously thought on the basis of palynological studies. According to this record, growing-season temperatures increased 6000-4500 yr B.P., decreased 4500-1500 yr B.P., and increased with fluctuations afterward. After 6000 yr B.P. stratigraphic changes in pollen percentages of Picea appear to be positively related with those of Mg/ Ca. This relationship implies that once the threshold of effective moisture was crossed for the establishment of Picea forests temperature was the primary control of Picea population density.
Bibliographical noteFunding Information:
This research was supported by the NSF Research Training Group on Paleorecords of Global Change at the University of Minnesota through a postdoctoral fellowship to FSH. Field work was funded by NSF Grant OPP-8922491, and laboratory analyses by NSF Grants ATM-9619583 and BIR-9014277 and NOAA Grant USDOC/NA 36GP0302. We thank Rick Knurr and Reed McEwan for laboratory assistance and Mary Edwards, Paul Glaser, Brian Haskell, and Herb Wright for comments. This paper is Limnological Research Center Contribution 503 and PALE Contribution 99.