Abstract
Major weather forecast centers produce physically based large-scale climate analyses and predictions that can be used as proxies for missing observations and thus as full-coverage climatologies. Because of this, a global reanalysis of recent climate is being carried out at the European Centre for Medium-Range Weather Forecasts (ECMWF). At the surface of the polar ice sheets (the atmospheric boundary condition for ice evolution), observations of climate are particularly scarce. To estimate how the new ECMWF climatology might help provide climate data over the polar ice sheets, the authors present 6 years of previously analyzed surface temperature and predicted precipitation for both Greenland and Antarctica. Analyses are the result of 6-h forecasts corrected to fit with reports from weather stations. Predicted variables are not corrected but the observation-constrained analyzed fields are used to initialize forecasting cycles. In spite of a sparse coverage of the observation network, the analyzed temperature, including seasonality, is very reasonable. Interannual variability, however, appears greater than suggested by satellite observation. Mean annual precipitation in Antarctica is fairly well represented, but it is difficult to determine whether a lack of seasonality on the plateau is reasonable or not. Precipitation in coastal Greenland is often too high, and accumulation might be low inland. Mean predicted accumulations, 1594 × 1012 and 539 × 1012 kg yr−1, over the Antarctic and Greenland ice sheets, respectively, are in good agreement with previous estimates. It is reasonable to expect that the reanalysis will largely satisfy the need for a full-coverage gridded climatology of the two polar ice sheets.
