Spatial and temporal variability of net snow accumulation over the Antarctic from ECMWF re-analysis project data

Abstract

Forecasts from the ECMWF re-analysis project (ERA) covering the period 1979–1993 are used to examine the spatial and temporal distribution of net snow accumulation (precipitation minus evaporation) over the Antarctic continent. There is generally good agreement between the spatial distribution of net accumulation in the model data, when the 15 year mean annual accumulation is considered, and the equivalent maps produced in earlier studies from in situ data. One of the major differences is the westerly displacement of the accumulation maximum on the western side of the Antarctic Peninsula as a result of the model orography having a high degree of smoothing in the east–west direction. The mean annual net accumulation in the ERA data for the whole of the continent is 151 mm year−1, equivalent to a total accumulation of 2106×1012 kg year−1. The mean accumulation value is in reasonable agreement with the best estimates from glaciological data, which recent studies have suggested is in the range 150–170 mm year−1. The lower value from ERA is partly a result of overestimation of evaporation/sublimation from the large ice shelves during the summer and spring, and an underestimation of the precipitation in the interior of the continent. The accumulation from the ERA data is the same as that computed from the operational ECMWF forecasts. During the 15 year data period, the mean accumulation varied from 129.1 mm (1987) to 171.8 mm (1984). Considering the continent as a whole, net accumulation was at a minimum in the summer season, although in coastal parts of West Antarctica, the minimum occurs in spring. The ERA accumulation data for the interior of the continent show no annual cycle and are significantly smaller than the available in situ measurements. At certain coastal sites in West Antarctica there is a clear relationship between annual precipitation and cyclone activity, although in East Antarctica such a relationship is only apparent in monthly data. Copyright © 1999 Royal Meteorological Society