Deriving Time Series of Ice-Sheet Accumulation Variations from Altimeter Measurements of Surface-Elevation Changes

Abstract:

Information on ice-sheet surface accumulation, A(t), is needed for studies of mass balance, polar climate variability, ice-sheet evolution, and dynamics. This paper describes a new method for deriving time series of A(t) variations, δA(t), on a monthly basis from continuous altimeter measurements. Changes in ice-sheet thickness and surface elevation are driven by a combination of δA(t), changes in the rates of firn compaction, and the rates of dynamic ice thickening or thinning (dHd/dt). Relevant time scales range from monthly fluctuations in δA(t) and surface temperature, Ts(t), to decadal to millennial scales of dynamic changes in ice flow. Our concept is based on the widely differing time-scales of principal processes, e.g. T> 5 years for dynamic changes and t = year/12 for δA(t) so that T >> t. An important aspect is our use of a baseline of ICESat-1 and ERS-1/ERS-2 altimeter measurements to obtain the multi-year (e.g. 5 to 17 years) dHd/dt of the Antarctic and Greenland ice sheets. We calculate changes in surface elevations caused by variations in the rate of firn compaction driven by both δA(t) and Ts(t) to adjust the altimeter-observed height changes, dH/dt, in order to derive dHd/dt over time periods of ≥ 5 years. Although monthly values of both δA(t) and Ts(t) are used, the derived dHd/dt is dependent only on the average change in δA(t) over the period and not on the higher-frequency monthly δA(t). On a monthly basis, the total accumulation-driven height change, dHaCA/dt, is a sum of the direct-height change, dHa/dt, and the change in firn compaction, dCA/dt, both caused by δA(t). dHa/dt is equal to δA(t)/ρs, where ρs is relative density of surface snow (typically 0.3) and δA(t) is w.e. Model simulations of dHaCA/dt driven by monthly δA(t) from ERA-interim re-analyses show that altimeter-observable dHaCA/dt and the desired dHa/dt = δA(t)/ρs are closely related. The relationship is better at colder locations such as the South Pole (Tm = -47̊C), for which the slope of dHaCA/dt versus dHa/dt is 0.989 and the correlation is 0.995 for 1982 to 2009, but it is also very good in warmer locations such as Law Dome (Tm = -17̊C), for which the slope is 0.816 and the correlation is 0.952. The relevant difference at South Pole is the slower response time of the firn compaction to perturbations.