Title: Aucun titre de diapositive
1Temporal and spatial structure of the Atlantic
eddy flow between 1980 and 2000 WOCE current
meter database versus 1/6 CLIPPER model
Thierry Penduff, Bernard Barnier, Christel Le
Delliou LEGI, Grenoble, France
APPROACH Observed EKEs Gather 1980-2000 WOCE
velocity data Selection / filtering /
interpolation EKE estimates over 6-months
intervals Model EKEs are Mirror images of
observed EKE estimates
OBJECTIVES Rigorously compare model and WOCE EKE
estimates over the Atlantic during 20 years
(Horizontal, vertical, and temporal structure)
- CLIPPER MODEL SIMULATION
- OPA 8.1 z-level code
- Primitive equations
- 68oS-65oN, 4 open boundaries
- 1/6o mercator grid, 42 levels
- 8-year spinup (Mean ERA15)
- 1980-2000 (ECMWF reanalysis
- analysis)
- WOCE CURRENT METER
- DATASET
- 1300 velocity time series
2Processing of the WOCE and CLIPPER databases
1300 raw WOCE velocity time series
CLIPPER 1980-2000 database
Quality tests gap filling or not, Tgt6 months, ...
Same locations and time intervals
31.5 rejected
891 velocity time series
5 day low-pass filter - Detrending - Variance
computation
891 WOCE and CLIPPER EKE estimates at the same
(x,y,z,t)
CLUSTERS
3Intensity of the mean EKE field within clusters
Gulf Stream Cluster Model EKE field is too
strong and baroclinic
4Baroclinicity of the mean EKE field
Baroclinicity B(EKEupper-EKElower)/EKElower
Shown Bmodel / BWOCE.
Model EKE too baroclinic
Model EKE too barotropic
Summary First systematic model/current meter
comparison study at basin scale Criteria
time-averaged EKE levels and baroclinicity In
most areas Model EKE levels somewhat too
weak (resolution) Model EKE too baroclinic
(spurious bottom friction in z-level
models) Model EKE level and baroclinicity
especially realistic in the NECC Future
Work Investigate the origin of problems Same
exercice on low-passed velocity timeseries