|
Titel |
The Bjerknes feedback in the tropical Atlantic from reanalysis and CMIP5 |
VerfasserIn |
Anna-Lena Deppenmeier, Rein Haarsma, WIlco Hazeleger |
Konferenz |
EGU General Assembly 2015
|
Medientyp |
Artikel
|
Sprache |
Englisch
|
Digitales Dokument |
PDF |
Erschienen |
In: GRA - Volume 17 (2015) |
Datensatznummer |
250104277
|
Publikation (Nr.) |
EGU/EGU2015-3694.pdf |
|
|
|
Zusammenfassung |
In this work we investigate the mechanism suspected behind the zonal mode of interannual
variability in the equatorial Atlantic: the Bjerknes feedback. The Bjeknes feedback is a closed
loop consisting of three interactions between the variable pairs named below, and is
known to be the driver of ENSO. A similar feedback loop is present in the tropical
Atlantic throughout the year, varying in strength on interannual time scales leading to
said mode of variability. We detect the components of this loop from reanalysis
(ERA-Interim and ORAS4) and compare them to the response seen in CMIP5 model
outputs.
From reanalysis the Bjerknes feedback can clearly be identified by correlating the three
relevant variable pairs: sea surface temperature anomalies (SST’) in the east of the basin with
zonal wind stress anomalies (Ïu’) across the equator; Ïu’ generated in the west of the
basin to heat content anomalies (HC’) most prominent in the east of the equatorial
Atlantic; and HC’ locally to overlying SST’. In the models, however, the feedback
cannot be identified quite as readily. While the influences of SST’ on Ïu’ and of Ïu’
on HC’ are included in the models, the interaction between HC’ and SST’ in the
eastern basin can hardly be seen. A shift in the seasonal cycle (of variance) reduce
correlation strengths in the model outputs for the first and the second part of the
loop, which can be corrected for by introducing a (one to two months) lag between
reanalysis and model. The representation of those parts of the feedback is, then,
satisfactory. The response of the SST’ to the underlying HC’, on the other hand, is and
stays weak with and without correction for seasonality. From reanalysis a clear
connection between these two variables can be found. The reason for the models’
discrepancies with observations most likely lies in the thermodynamic influence
on the SSTs, which overshadows the heat content’s influence. This is in addition
to the fact that the simulated thermocline in the east of the equatorial Atlantic is
too deep to efficiently cool SSTs in the cold tongue region, leaving it too warm. |
|
|
|
|
|