An examination of the precipitation regime at Thurston Island, Antarctica, from ECMWF re-analysis data

Abstract

In this study numerical weather prediction model data are utilized to examine the precipitation regime at Thurston Island (TI) (∼72°S, 99°W) in West Antarctica. This region was chosen because the precipitation may well reflect the high variability of cyclones in the Amundsen Sea, which in some years appears related to the El Nino–Southern Oscillation (ENSO) phenomenon. The data used are derived from the European Centre for Medium-range Weather Forecasts (ECMWF) Re-Analysis (ERA) project, encompassing the 15-year period from 1979 to 1993. A comparison of monthly TI precipitation derived from both ERA and ECMWF operational data for a 101-month overlap period demonstrates that the known ERA problems associated with moisture fluxes over Antarctica are not significant at TI. The annual precipitation cycle at TI (mean of 713 mm water equivalent) is related directly to the frequency of cyclones in the Amundsen Sea, which, in turn, reflects the semi-annual oscillation. The majority of these systems develop in the circumpolar trough although ∼4% undergo cyclogenesis east of New Zealand. No trends in precipitation at TI can be discerned in the model data. Significant precipitation at TI occurs when the longwaves over the Pacific are amplified such that an intense low (blocking high) pressure is located west (east) of the island. Correlations between TI precipitation and mean sea level pressure in these two regions suggest that they are of equal importance in determining the precipitation that falls at TI. There is a small but significant anticorrelation between TI precipitation and the Southern Oscillation Index (SOI) from 1990 onwards. It is postulated that an increase in cyclone intensity during the El Nino years of the early 1990s is at least partially responsible because cyclone numbers are well correlated to the SOI during this period. No consistent relationship between TI precipitation and SOI is observed prior to 1990 in the ERA data. Copyright © 2000 Royal Meteorological Society