Climate reconstruction using data assimilation of water isotope ratios from ice cores

Water isotope data from ice cores, particularly δ¹⁸O, have long been used in paleoclimatology. Although δ¹⁸O has been primarily interpreted as a proxy for local air temperature, isotope-enabled climate models have established that there are many nonlocal and nontemperature-related climatic influences on isotopic signals at coring locations. Moreover, recent observational studies have linked ice core isotopes to nonlocal patterns of climate variability, particularly to midlatitude atmospheric circulation patterns and to variations in tropical climate. Therefore, paleoclimate reconstructions may better utilize ice core isotope proxies by combining them with isotope-enabled climate models. Here we employ a data assimilation-based technique that fuses isotopic proxy information with the dynamical constraints of climate models. Through several idealized and real proxy experiments we assess the spatial and temporal extent to which isotope records can reconstruct surface temperature, 500 hPa geopotential height, and precipitation. We find local reconstruction skill to be most robust across the reconstructions, particularly for temperature and geopotential height, as well as limited nonlocal skill in the tropics. These results are in agreement with long-held views that isotopes in ice cores have clear value as local climate proxies, particularly for temperature and atmospheric circulation. These results also show that in principle nonlocal climate information may also be inferred from ice cores. However, the spatial range of this information is nonuniform and depends on skillful modeling of the proxy data within the reconstruction process.