Physically based summer temperature reconstruction from ice layers in ice cores.

Previous studies reconstructing summer temperature from an ice core relied on statistical relationship between melt feature and instrumental temperature observed at a nearby station. This study demonstrates a novel method to reconstruct summer temperature from ice layer thickness in an ice core using an energy balance model, in which heat conduction and refreezing of meltwater in firn are taken into account. Using seasonal patterns of the ERA-Interim reanalysis dataset for an ice core site, we calculated amounts of refreezing water within firn under various settings of summer mean temperature (SMT) and annual precipitation, and prepared lookup tables containing these three variables. We then estimated SMT from the refreezing amount and annual accumulation, both available in an ice core. We applied this method to four ice cores drilled in the sites of different climates; two sites on the Greenland Ice Sheet, one in Alaska, and one in Russian Altai Mountains. Reconstructed SMTs show comparable variations to the temperatures observed at nearby stations. Relationships between SMT and ice layer thickness differ site by site, indicating that a single approximation cannot be applicable to estimate SMT. Sensitivity analyses suggest that annual temperature range, amount of annual precipitation and firn albedo setting significantly affect the relationship between SMT and ice layer thickness. This new method provides alternative and independent estimation of SMT from ice cores affected by melting. [ABSTRACT FROM AUTHOR]