A composite isotopic thermometer for snow

[1] The isotopic δ thermometer for snow is an equation that relates the isotopic δ value (derived from 18O/16O or D/H values) to some prescribed reference temperature. In contrast to the usual linear least squares fit of mean annual isotopic 〈δ〉 versus site mean annual surface temperature 〈Ts〉, a cubic equation is fitted to a composite, global, spatial (〈δ〉, 〈Ts〉) data set. A ±4‰ local variance along the curve indicates that δ values are influenced by factors other than the temperature signal, but in the long-scale smoothing process these factors are filtered out. As 〈Ts〉 data over the complete range of raw temperatures are inhomogeneous, because of the existence of strong temperature inversions in the cold half, a transformation scheme is applied to homogenize the data. This produces a new series, 〈Tatm〉. These values are more likely to be near the mean condensation temperature (〈Tc〉) of the snow. The slope, d〈δ〉/d〈Tatm〉, as a quadratic function of 〈Tatm〉, best fits many published spatial and temporal slope determinations. The results have obvious application to the theory of isotopic processes not covered here. However, in an important practical application, the new equation can be utilized in the construction of diagrams of δ versus altitude where a δ value needs to be obtained from a known or calculated air temperature and where field data are gapped, especially at high altitude. Such diagrams can be used to demonstrate cases where site-specific δ thermometers may become invalid because of atmospheric structural changes, as shown in an associated paper that will be briefly outlined.