and Cl and other bulk-incompatible elements in
all of these sections, which requires apatite in the
fractionating assemblage and is not consistent
with an assemblage consisting entirely of NAMs.
Our model also predicts that residual melts
should be poor in elements such as F and P that
are highly compatible in apatite and are thereby
depleted during apatite fractionation. An example
is the glass in contact with apatite from 10044,644
discussed above (16). This highly evolved glass
(>78% SiO2) has low F, Cl, and P (F < 460 ppm,
Cl < 90 ppm, and P < 350 ppm) and yet makes up
only 0.1 T 0.06% of the total volume of 10044
(26). This is consistent with depletion of the melt
in apatite components far below the thousandfold
increase that might be expected for a late-stage
melt refined by a high degree of fractional crystallization of NAMs. Because fractional crystallization models indicate that order-of-magnitude
or greater increases in H2Oap are possible, it is not
trivial to use apatite to estimate minimum water
contents in lunar magmas. In cases where fractionation of apatite is a concern, the lowest H
values are likely the ones least affected by loss of
F and Cl and make for the most robust interpretation of H2Omelt, but only as a maximum
value. These H-enriched apatites may be quite
abundant: A melt with H2O = F = Cl = 1 ppm bulk
(10 ppm at apatite saturation) will produce apatite
with initial H2O abundances of ~200 ppm. This
is also the abundance that would be predicted for
all apatite crystallizing under conditions approximating equilibrium crystallization. However,
given the same starting conditions, fractional crystallization models produce more than 40% by
mass apatite with >1000 ppm H2O and more than
20% apatite with more than 10,000 ppm H2O.
Apatite has such a strong effect on the volatile
content of a magma during fractional crystalliza-
tion that it is difficult, if not impossible, to learn
anything about what conditions were like before
apatite saturation unless equilibrium conditions
are maintained throughout the life of the apatite.
If only H, F, and Cl are considered, the patterns of
degassing [decrease in the H/(H + F + Cl) content
of apatite with time] and fractionation [increase
of the H/(H + F + Cl) content of apatite with time]
can be indistinguishable, and other parameters
(such as nonvolatile trace elements) should be
used to determine the forward-time direction of
apatite evolution. Because it is not required that
late-stage H2Omelt be elevated in order to explain
the elevated abundances of H2Oap, hydrogen-rich
apatite cannot be cited as evidence for elevated
H2Omelt a priori. This permits reconciliation of
H-rich apatites with the high temperatures and
associated volatile depletion required by the giant-
impact hypothesis (9) without the need to find those
regions of impact-model space that are most per-
missive to retention of volatiles. Our results are
also consistent with the nominally anhydrous 37Cl-
enrichment mechanism proposed for lunar mate-
rials by (8). Apatite can still serve as a useful target
for isotopic studies of H and Cl provided that
those studies take into account the possibility that
multiple, initially low-H sources (such as solar
wind, surface ice, and spallation products) might
be amalgamated and concentrated into apatite. Ex-
cluding apatite, the geochemical evidence for large
portions of the interior of the Moon being more
than nominally anhydrous is greatly reduced.
References and Notes
1. S. R. Taylor, C. M. Pieters, G. J. MacPherson, Rev. Mineral.
Geochem. 60, 657–704 (2006).
2. J. P. Greenwood et al., Nat. Geosci. 4, 79–82 (2011).
3. F. M. McCubbin et al., Proc. Natl. Acad. Sci. U.S.A. 107,
4. J. W. Boyce et al., Nature 466, 466–469 (2010).
5. E. H. Hauri, T. Weinreich, A. E. Saal, M. C. Rutherford,
J. A. Van Orman, Science 333, 213–215 (2011).
6. R. Tartèse, M. Anand, Earth Planet. Sci. Lett. 361,
7. J. J. Barnes et al., Chem. Geol. 337-338, 48–55
8. Z. D. Sharp, C. K. Shearer, K. D. McKeegan, J. D. Barnes,
Y. Q. Wang, Science 329, 1050–1053 (2010).
9. W. K. Hartmann, D. R. Davis, Icarus 24, 504–515 (1975).
10. H. Hui, A. H. Peslier, H. Zhang, C. R. Neal, Nat. Geosci. 6,
11. K. Pahlevan, D. J. Stevenson, Earth Planet. Sci. Lett. 262,
12. J. D. Webster, C. Tappen, C. Mandeville, Geochim.
Cosmochim. Acta 73, 559–581 (2009).
13. E. Mathez, J. Webster, Geochim. Cosmochim. Acta 69,
14. M. Huh, J. Boyce, C. Manning, E. Medard, Am. Geophys.
Union Fall Meeting Abst. 1, 2534 (2011).
15. K. Vander Kaaden, F. McCubbin, E. Whitson, E. Hauri,
J. Wang, Proc. Lunar Planet. Sci. Conf. 43, abstr. 1247
16. Methods and materials, figs. S1 to S5, and tables S1 and
S2 are found in the supplementary materials.
17. E. H. Hauri, G. A. Gaetani, T. H. Green, Earth Planet. Sci.
Lett. 248, 715–734 (2006).
18. J. A. O'Leary, G. A. Gaetani, E. H. Hauri, Earth Planet. Sci.
Lett. 297, 111–120 (2010).
19. M. M. Hirschmann, Annu. Rev. Earth Planet. Sci. 34,
20. C. Beyer, S. Klemme, M. Wiedenbeck, A. Stracke,
C. Vollmer, Earth Planet. Sci. Lett. 337–338, 1–9 (2012).
21. A. Fabbrizio, R. Stalder, K. Hametner, D. Günther, Geochim.
Cosmochim. Acta 121, 684–700 (2013).
22. J. W. Boyce, J. M. Eiler, M. C. Channon, Am. Mineral. 97,
23. P. Roeder, R. Emslie, Contrib. Mineral. Petrol. 29,
24. M. J. Toplis, Contrib. Mineral. Petrol. 149, 22–39
25. M. Anand, R. Tartese, J. J. Barnes, I. A. Franchi, N. A. Starkey,
Mineral. Mag. 77, 589 (2013).
26. D. Beaty, A. Albee, Proc. Lunar Planet. Sci. Conf. 9,
27. M. Anand et al., Geochim. Cosmochim. Acta 70,
28. S. M. Elardo et al., Meteorit. Planet. Sci. 49, 261–291
29. F. McCubbin et al., Am. Mineral. 95, 1141–1150 (2010).
30. F. M. McCubbin et al., Geochim. Cosmochim. Acta 75,
Acknowledgments: J. W.B. acknowledges support from NASA
Cosmochemistry grant NNX13AG40G; A.H. T. and J. W.B.
acknowledge support from NASA Cosmochemistry grant
NNX12AH64G to A.H. T. F.M.M. acknowledges support from
NASA Lunar Advanced Science for Exploration Research grant
NNX13AK32G. J.P.G. acknowledges support from NASA
Lunar Advanced Science for Exploration Research
NNX11AB29G and J. Eckert (Yale) for assistance with the
electron-probe microanalysis. All data used in the modeling
and testing of the model for this paper are available in the
supplementary materials or in the published, refereed journal
articles that are cited in the references section. J. W.B.,
F.M.M., and A.H. T. conceived of the solution to the lunar
apatite paradox. S.M. T. and J. W.B. wrote the numerical model.
J.P.G. provided essential tests of the model. All authors
contributed to the writing of the paper.
Materials and Methods
Figs. S1 to S5
Tables S1 and S2
3 January 2014; accepted 6 March 2014
Published online 20 March 2014;
Preservation of a Preglacial
Landscape Under the Center of the
Greenland Ice Sheet
Paul R. Bierman,1 Lee B. Corbett,1 Joseph A. Graly,1† Thomas A. Neumann,1‡ Andrea Lini,1
Benjamin T. Crosby,2 Dylan H. Rood3,4§
Continental ice sheets typically sculpt landscapes via erosion; under certain conditions, ancient
landscapes can be preserved beneath ice and can survive extensive and repeated glaciation.
We used concentrations of atmospherically produced cosmogenic beryllium-10, carbon, and
nitrogen to show that ancient soil has been preserved in basal ice for millions of years at the
center of the ice sheet at Summit, Greenland. This finding suggests ice sheet stability through
the Pleistocene (i.e., the past 2.7 million years). The preservation of this soil implies that the
ice has been nonerosive and frozen to the bed for much of that time, that there was no substantial
exposure of central Greenland once the ice sheet became fully established, and that preglacial
landscapes can remain preserved for long periods under continental ice sheets.
Adiverse set of geochemical records has been developed from ice recovered in the 3054-m Greenland Ice Sheet Project 2 (GISP2) core. These data provide a detailed his- tory of climate and ice dynamics stretching back over 100,000 years. The lowermost 13 m of the