Which reanalysis data is good for momentum diagnostics of extratropical stratosphere?

Wednesday, 17 December 2014
Seok Woo Son, Seoul National University, Seoul, South Korea, Patrick Martineau, McGill University, Montreal, QC, Canada and Masakazu Taguchi, Aichi Univ of Education, Kariya, Japan
Dynamic consistency of eight reanalysis datasets is evaluated in terms of momentum budget of polar stratospheric vortex. Their consistency is quantified by computing the residue of 6-hourly zonal-mean wind tendency that is not explained by zonal-mean momentum equation. Both Eulerian-mean and transformed Eulerian-mean equations are considered. In all datasets, absolute residues rapidly increase with heights partly due to unresolved wave forcings (e.g., gravity waves) in momentum equations. However, in comparison to early generation datasets, modern reanalysis datasets generally show an improved dynamic consistency at all levels. The improvement from NCEP-NCAR to NCEP-DOE and to NCEP-CFSR reanalysis data is very discernible. Similar improvement is also found from ERA-40 to ERA-Interim, and from JRA-25 to JRA-55 data.

The residues and individual forcing terms of momentum equations are further evaluated in the course of the vacillation cycle of the stratospheric polar vortex, i.e., quasi-periodic intensification and weakening of the wintertime stratospheric winds, to identify the source of the residues and their inter-data differences. The residues vary in function of the vacillation cycle with relatively small residues in weak and strong vortex states but large residues in intermediate state, indicating that dynamic consistency of reanalysis data may depend on the mean state. Throughout the vacillation cycle, the largest discrepancy among datasets is found in the Coriolis torques acting on the Eulerian-mean circulation, explaining a large portion of inter-data spread of the residues. Transformed Eulerian-mean diagnostics of wave-mean flow interaction, such as the EP-flux divergence, are generally consistent across reanalysis dataset. This indicates that a good representation of mean meridional circulation is crucial for dynamic consistency of stratospheric wind variability in the reanalysis data.