Ice/firn age distribution on the Elbrus Western Plateau (Caucasus) inferred from ice flow model

Abstract. The glaciers of Mount Elbrus (Caucasus) contain paleoclimatic and paleoenvironmental information representative of a vast region. The cold conditions and negligible seasonal melting in the near-summit area of Elbrus ensure the excellent preservation of climatic signals. In 2009, a 182.65 meter long ice core was obtained from the Western Plateau (WP) of Elbrus at approximately 5100 ma.s.l. This core was partially dated using chemical stratigraphy (upper part) and radiocarbon dating (basal samples), serving as the basis for numerous paleoreconstructions. However, the ice age distribution within the intermediate part of the core and across the entire glacier on the WP remained unknown. In this work, a three-dimensional steady state full Stokes flow model for a cold glacier with a rheological law accounting for firn densification was applied both in a purely mechanical and in a thermomechanically coupled versions. Using this model, the ice velocity field was simulated in the central part of the WP. Ice age distribution was determined by solving a boundary value problem for the dating equation. All calculations were performed using Elmer/Ice, a finite element software designed for ice dynamics modeling. The model was calibrated by fitting the simulated age-depth relationship to the observed data from the ice core, with ice viscosity used as the primary calibration parameter. This approach provided a series of three-dimensional ice age distributions on the WP for various modeling scenarios. All versions of the model accurately reproduce the ice age according to ice core data to a depth of 140–150 m (130–180 years). Below 150 m, the ice age increases sharply and the dating discrepancies between different modeling scenarios become larger. Overall the modelled ages fell within 68.2 % confidence intervals of the radiocarbon dated near-bottom ice samples which indicated mean radiocarbon ages ranging from 1 to 2 ka. However, the model was unable to resolve the dating of the basalice layer up to 3–4 m thick. Future model improvements should focus on refining basal conditions, including accounting for potential melting, and identifying areas containing the oldest ice.