Validity of the temperature reconstruction from water isotopes in ice cores.

Journal of Geophysical Research, Vol. 102, No. C12, p. 26,471-26,487, 1997

J. Jouzel
Laboratoire de Modélisation du Climat et de l'Environnement , Gif-Sur-Yvette, France.
R. Alley
Earth system Science Center and Department of Geosciences, Pennsylvania State University University Park.
K.M. Cuffey
Department of Geological sciences and Quaternary Research Center, Washington University Seattle.
W. Dansgaard, S.J. Johnsen
Departement of Geophysics, The Niels Bohr Institute of Astronomy, Physics and Geophysics, University of Copenhagen.
P. Grootes, M. Stuiver
Department of Geological Sciences and Quaternary Research Center, University of Washington, Seattle.
G. Hoffmann
Max Planck-Institut für Meteorologie, Hamburg, Germany .
R.D. Koster
Hydrologic Science Branch, NASA Goddard Space Flight Center, Greenbelt Maryland.
D. Peel
British Antarctic Survey, Natural Environment Research Council, Cambridge, England.
C.A. Shuman
Ocean and Ice Branch, NASA Goddard Space Flight Center, Greenbelt Maryland.
M. Stievenard
Laboratoire de Modélisation du Climat et de l'Environnement, Gif-Sur-Yvette, France.
J. White
Institute of Alpine and Arctic Research and Department of Geological Sciences, University of Colorado, Boulder.

Well-documented present-day distributions of stable water isotopes (HDO) and H2 18O show the existence, in middle and high latitudes, of a linear relationship between the mean annual isotope content of precipitation (δD and δ18O) and the mean annual temperature at the precipitation site. Paleoclimatologists have used this relatiionship, which is particularly well obeyed over Greenland and Antarctica, to infer paleotemperatures from ice core data. There is, however, growing evidence that spatial and temporal isotope/ surface temperature slopes differ, thus complicating the use of stable water isotopes as paleothermometers. In this paper we review empirical estimates of temporal slopes in polar regions and relevant information that can be inferred from isotope models: simple, Reyleigh-type distillation models and (particularly over Greenland) general circulation models (GCMs) fitted with isotope tracer diagnostics. Empirical estimates of temporal slopes appear consistently lower than present-day spatial slopes and are dependent on the timeclock considered. This difference is most probably due to changes in the evaporative origins of moisture, changes in the seasonality of the precipitation, changes in the strength of the inversion layer, or som combination of these changes. Isotope models have not yet been used to evaluate the relative influences of these different factors. The apparent disagreement in the temporal and spatial slopes clearly makes calibrating the isotope paleothermometer difficult. Nevertheless, the use of a (calibrated) isotope paleothermometer appears justified; empirical estimates and most (though not all) GCM results support the practice of interpreting ice core isotope records in terms of local temperature changes.