A comparison of reanalyses in the tropical stratosphere. Part 2: the quasi-biennial oscillation

Abstract

 Reanalysis datasets potentially offer the opportunity to examine the tropical quasi-biennial oscillation (QBO) in greater detail than in the past, including the associated meridional circulation and the links with other parts of the atmosphere. For such studies to be useful, the QBO represented by the reanalyses should be realistic. In this work, the QBO in the ERA and NCEP reanalyses is validated against rawinsonde observations from Singapore. Monthly mean data are used. In the lower stratosphere (at 50 hPa and 30 hPa) the ERA QBO is reasonable, although the wind extrema in both phases are too weak and the vertical shear and the temperature anomalies are too small. The NCEP QBO is weaker still. At 10 hPa neither reanalysis system performs well, both systems failing to reproduce the westerlies, possibly because of the proximity of the upper boundary. The Singapore wind is representative of the zonal means in the reanalyses. The weak wind extrema in the reanalyses would not support a wave-mean flow interaction theory of the QBO, because a large portion of the gravity wave spectrum which would be absorbed in reality would be transmitted beyond 10 hPa. The stronger shear zones captured in the ERA data are associated with larger, more realistic temperature perturbations near 30 hPa. The northward velocities in the NCEP data show a more realistic structure than in the ERA reanalysis, where they are dominated by a vertical “gridpoint wave” structure in the lowermost stratosphere. Despite the shortcomings of the reanalyses, the high correlations of the wind at 30 hPa and 50 hPa with the observations at Singapore mean that the reanalyses could potentially be used to examine the effects of the QBO away from the tropical stratosphere. Future reanalyses need to take full account of the wind shears evident in the rawinsonde observations and use models with an adequate resolution to capture these vertical scales.