Uncertainties in atmospheric surface pressure fields from global analyses

Abstract

Operational analyses from the U.S. National Centers for Environmental Prediction (NCEP) and the European Centre for Medium‐Range Weather Forecasts (ECMWF) provide, among other quantities, atmospheric surface pressure. This field is especially useful for data processing and interpretation of altimeter and gravity satellite missions like Jason‐1 and GRACE (Gravity Recovery and Climate Experiment). Quantitative use of this pressure field, however, requires knowledge of its uncertainties. To assess these uncertainties, we compare NCEP and ECMWF analyses of surface pressure to an extensive set of surface barometric observations for the period 2001–2005, both land‐ and ocean‐based, as well as to each other. Differences between each analysis and the observations are quite similar, indicating that both analyses generally fit the station observations equally well. Such differences, including both bias and variance about the bias, are largest over parts of the Southern Ocean, and over Asia, central Africa, and Antarctica. These bias and variance differences between the NCEP and ECMWF surface pressure fields change only slightly over the period of study. More importantly, though, such differences are substantially smaller than those found by previous studies for an earlier period (1993–1995), suggesting recent improvements in both analyses. The largest differences are found over high latitudes, particularly over Antarctica, and in the respective winter hemisphere. Much of the differences occur on rapid timescales, though with a significant portion at timescales longer than 1–2 months. The potential impact of such uncertainties on de‐aliasing and interpreting sea level and gravity data from altimeter and GRACE missions is discussed.