Cold-Point Tropopause Temperature Bias in Reanalyses Modulated By Equatorial Waves

In this reanalysis inter-comparison study, we compare the modeled Cold-point tropopause temperature (CPT from model level data) near the equator to that of multi-mission GNSS-RO within a preliminary time period of 2007-2011. The known reanalysis CPT temperature warm bias is studied relative to the phase of equatorial waves filtered from GNSS-RO at tropopause level. This uncovers a common denominator in the CPT temperature bias behavior in all reanalyses:

- When equatorial waves are present, within their cold phase the reanalysis bias is increased by over 1K, i.e. >1K on top of the reanalysis average warm bias.

- Within an equatorial wave’s warm phase, the opposite is observed, with the reanalysis CPT temperature bias even being reversed at higher equatorial wave amplitudes, i.e. the modeled CPT becoming too cold in these cases.

- With no equatorial wave anomalies present, the reanalysis bias remains at the average reported in previous inter-comparisons that looked at zonal-mean CPT biases.

Inertia-Gravity and Mixed Rossby-Gravity waves show this behavior most clearly, Kelvin waves show a similar modulation but with half of the magnitude, and equatorial Rossby wave modulation of CPT reanalysis bias is still apparent but with further decreased magnitude. This hints on a time-space scale-dependence, with medium-smaller-scale and faster equatorial waves influencing CPT reanalysis bias the most. For comparison, the large-scale and slow-moving MJO does not show any of the effects described above.

Current work focuses on the data assimilation response of the reanalysis systems, again relative to the equatorial wave’s phases.