The representation of the Quasi-Biennial Oscillation (QBO)
and Semi-Annual Oscillation (SAO) in the IFS depends on
the model formulation in the tropical region, particularly
the parameterisation of the effects of gravity waves.
Therefore, the representation of the QBO and SAO vary with
model cycle and resolution. Observations do provide
constraints for the analyses, but poorly observed regions,
such as the upper stratosphere and the mesosphere,
will exhibit similar characteristics to those of the forecast
model.
The depiction of the QBO in the lower to mid stratosphere
(from about 50 to 5 hPa) in ERA5 is very similar to that in
ERA-Interim (Figure \ref{fig:qbo}a and b), at least for the period considered
here, 2008 to 2017. The similarity is present in both the
strength of the descending easterlies and westerlies and in
their phase and periodicity. The disruption of the westerlies
just above 50 hPa in early 2016 is clearly visible in both
reanalyses.
Although the broad representation of the SAO between 5 and
0.5 hPa is reasonably similar in the two reanalyses, with
descending easterlies and westerlies at similar times of the
year, the magnitude of the winds and the exact pattern of
descent are not the same (Figure \ref{fig:qbo}c and d). However, between 0.5 and
0.1 hPa the agreement between the reanalyses is poor, with
the two representations of the SAO being very different. In
this region, ERA5 has westerlies that are up to 30 ms-1
larger than those in ERA-Interim and ERA5 also has a complete
absence of descending easterlies. This predominance of
westerlies in ERA5 is related to the spurious equatorial
mesospheric jet that occurs in CY41R2 of the IFS and which
peaks in the transition seasons.