To investigate the history of Pleistocene lake levels in the southwestern U.S.A., and to improve our understanding of the systematics of in situ produced cosmogenic nuclide accumulation in hot climates and young rocks, we examined the 3He, 10Be and 26Al contents of quartz separates from metamorphic quartzites and quartz-rich felsic rocks incorporated in a series of stranded conglomerate beach ridges ranging from 3 to 160 m above sea-level in Death Valley, California. Calculated 10Be ages range from 17 to 135 kyr, with most ages between 20 and 60 kyr. In contrast, cosmogenic 3He contents indicate exposure ages of < 1 kyr, except for one sample at 18 ± 4 kyr for which the 10Be age is still significantly greater (44 ± 11 kyr). 26Al 10Be ratios close to the cosmogenic production ratio rule out the presence of meteoric 10Be and point to 3He loss as the cause of the age discrepancies. Comparison to laboratory diffusive loss measurements using a diurnal direct solar heating model implies the 3He loss occurs because the effective quartz grain sizes are < 100 μm - far smaller than the 1-3-mm fracture spacings estimated from thin sections. Because direct insolation enhances 3He loss, sampling from shallow depths (∼20-40 cm) may be useful in some cases. 4He and 20Ne measurements demonstrate that small amounts of radiogenic and atmospheric 3He are also present at levels equivalent to a few kiloyears sea-level exposure, which should be taken into account when dating surfaces of similar age. The lack of any clear progression with terrace altitude and the occurrence of two-fold variations in apparent exposure ages within individual terraces suggest that the 10Be contents do not solely represent accumulation in the present sample locations. Instead, accumulation probably also occurred during sample exposure in the higher-altitude beach rock source regions and during fluvial transport to the ancient lake shore. Thus the beach terrace exposure ages of ∼20-60 kyr are probably maximum ages for the times of these high lake level stands in Death Valley.
Bibliographical noteFunding Information:
This project was initiated through discussions and a short field trip with B.C. Burchfiel, J.M. Edmond and P. Molnar. The solar heating model was calculated using MATLAB@ software from The Mathworks, Inc. Special thanks to David Graham (OSU ) for performing the high-temperature extraction to verify complete 3 lease, to D. DebofIle and J. Lestringuez consistent quality of their work at the Tande-tron, and to E.J. Brook (URI) for the pre-publication estimate of the 3He/‘oBe cosmogenic production ratio in Antarctic quartzites. Helpful comments from Mark Kurz and two anonymous reviewers were appreciated. T.W.T. acknowledge support from INSU, DBT Theme 1: Fleuves et erosion. E.T.B. acknowledges support from NSF through grant INT-900864 I, a NATO Postdoctoral Fellowship, and the INSU DBT program. Tandetron operation is supported by the CNRS, CEA and IN2P3.