during snow formation to generate an ice isotopic ratio signal. Both of these factors increase in strength as temperature is reduced and, consequently, the correlation of the ice isotopic ratio with temperature is very strong in the cold interiors of the polar ice sheets. Ice isotopic ratio records are geographically limited to locations of significant ice thickness, namely, polar regions and high-altitude mountain ranges elsewhere. It is remarkably fortuitous that the high Andes, the Tibetan and Himalayan ranges, and the great volcanoes of eastern equatorial Africa offer any ice records at all for Earth’s low latitudes.
The temperature dependence of the ice isotopic ratio arises from fundamental physics at the molecular scale combined with geophysical processes at the planetary scale (Dansgaard 1964, Kavanaugh and Cuffey 2003). However, additional influences on ice isotopic ratio can be significant (Dansgaard 1964, Pierrehumbert 1999, Alley and Cuffey 2001, Kavanaugh and Cuffey 2003, Jouzel et al. 1997), and, consequently, ice isotopic ratio measurements must be calibrated against independent temperature information in order to be used as a quantitatively accurate thermometer (Cuffey et al. 1995). Such calibrations have been applied for long-timescale records from the polar ice sheets, but not for low-latitude high-altitude ice cores (from the Andes, Kilimanjaro, and Tibet), where it is more difficult to isolate and quantify the temperature component of the signal. In general, the ice isotopic ratio records from the interior regions of polar ice sheets yield good temperature reconstructions (Alley and Cuffey 2001, Cuffey et al. 1995). The low-latitude ice isotopic ratios yield a climate signal that depends on a variety of hydrologic and thermal influences in the broad geographic region that supplies moisture to the high glaciated mountains (Pierrehumbert 1999, Tian et al. 2003, Vuille et al. 2003a, Hoffmann et al. 2003, Thompson and Davis 2005, Alley and Cuffey 2001, Jouzel et al. 1997). The connection of ice isotopic ratio to temperature becomes stronger at lower temperatures (e.g., Kavanaugh and Cuffey 2003, Jouzel et al. 1997).
All glacial sites in the low latitudes are cold enough that this temperature influence will have some reflection in the ice isotopic ratio (Pierrehumbert 1999, Tian et al. 2003). Ambiguity results from significant residual influences of warmer regions upwind and local processes related to snowfall timing and preservation (Hardy et al. 2003). Precipitation amount is the effect of greatest importance (Dansgaard 1964). Near sea level at low latitudes the isotopic ratios in precipitation are demonstrably not reflective of temperature changes at ground level—the water distillation process primarily happens in the vertical dimension in large storm clouds, resulting in a correlation between precipitation amount and isotopes at ground level, rather than a correlation between temperature and isotopes. The high-altitude low-latitude ice core sites are in a transition zone where both temperature and this precipitation effect have influence (Pierrehumbert 1999).
In Tibet, ice isotopic ratios in the south appear to be dominantly influenced by monsoonal precipitation, whereas in the north, temperature dominates (Yao et al. 1996, Tian et al. 2003). In the equatorial Andes, ice isotopic ratios retain a strong influence of precipitation over the Amazon lowlands and partly correlate with both Pacific sea surface temperatures and Amazonian temperatures (Henderson et al. 1999,