• Keine Ergebnisse gefunden

2 Inverse 2D ocean model: FEMSECT

N/A
N/A
Protected

Academic year: 2022

Aktie "2 Inverse 2D ocean model: FEMSECT"

Copied!
1
0
0

Wird geladen.... (Jetzt Volltext ansehen)

Volltext

(1)

Stiftung Alfred-Wegener-Institut für Polar- und Meeresforschung

75oW 50oW 25oW 0o 15oN

30oN 45oN

60oN

75oN

[m]

−1 −0.5 0

The picture shows the geodetic Rifu- gio MDT from section 1 before mod- ification by the any ocean model.

The Rifugio MDT is now combined with the ocean model IFEOM.

The inverse error covariance matri- ces are used for weighting the data.

1 Mean dynamic topography modeling

Mean dynamic topography (MDT) η is the departure of the sea surface from the geoid. The geostrophic relation balance

v = g f

∂η

∂x

is used for deriving ocean surface velocities v. These can be assimilated into inverse ocean models.

satellite orbit sea surface

η

geoid reference ellipsoid

Mean dynamic topography (MDT)

Common problems of MDT models are the different scales, accuracies and basis functions of altimetry data and geoid data.

These issues are avoided by combining the normal equations directly. The MDT is calculated on the ocean model grid and no additional smoothing of the MDT is required.

A stochastic model accounts for the omission error in the respective frequency domains.

GRACEAltimetrySmoothness

=

=

=

Combination

=

Elimination of xcs

Normal equation for MDT

Integration

into ocean model

xcs1 xcs2 xcs3 xFE

Frequency domains of observations and parametrization

Combining

mean dynamic topography (MDT) and steady-state ocean models

Grit Freiwald1, Martin Losch1, Wolf-Dieter Schuh2 and Silvia Becker2

1Alfred Wegener Institute for Polar and Marine Research, Bremerhaven, Germany

2University Bonn, Institute of Theoretical Geodesy, Germany Email: grit.freiwald@awi.de

2 Inverse 2D ocean model: FEMSECT

FEMSECT describes a 2D section from Tasmania to Antarctica.

The full combined Rifugio approach is not applicable to a 2D section model. However, the effects of different omission error models are studied.

130

oO

140oO 150oO 160oO

170o

O

70oS 60oS 50oS 40oS 30oS

SAF COUN

PF

EAC

Section location. latitude [°S]

depth [m]

Across−section velocities, positive eastward [m/s]

45 50 55 60 65

−4500

−4000

−3500

−3000

−2500

−2000

−1500

−1000

−500 0

−0.4

−0.3

−0.2

−0.1 0 0.1 0.2 0.3 0.4

Velocity profile across the section.

The velocity profile across the section is already well repre- sented without additional MDT information. When MDT is included, the transport error estimates change.

FEMSECT transport errors:

0 Ref.

5 10 15 [Sv]

0 0.1 0.2 0.3

[PW]

−−−−−−−−−−−−−−−−−→

increasing consideration of omission error

Final transports across section:

Mass: 159 ± 17 Sv Heat: 1.91 ± 0.26 PW

4 Results

FEMSECT:

• Additional MDT information reduces the transport error.

• Neglecting the omission error underestimates the true error.

IFEOM:

• Model-data deviations are consistent with error estimates.

• Estimation of posterior errors is not straightforward and a sub- ject for further research.

• The geodetic MDT is improved by the ocean model.

BLABLA BLABLA BLABLA BLABLA

• The ocean model is improved by the new MDT:

– more detailed current structure

– steeper temperature gradients

– more realistic MOC

– increased heat transports

3 Inverse 3D ocean model: IFEOM

IFEOM is a 3D Finite Element model for the North Atlantic ocean.

Without Rifugio MDT: With Rifugio MDT:

75oW 50oW 25oW 0o 15oN

30oN 45oN

60oN

75oN

[m]

−0.5 0 0.5

75oW 50oW 25oW 0o 15oN

30oN 45oN

60oN

75oN

[m]

−0.5 0 0.5

Resulting IFEOM MDT.

75oW 50oW 25oW 0o 15oN

30oN 45oN

60oN

75oN

[Sv]

−50 0 50

75oW 50oW 25oW 0o 15oN

30oN 45oN

60oN

75oN

[Sv]

−50 0 50

IFEOM barotropic stream function.

[m]

20oN 40oN 60oN 80oN

−5000

−4000

−3000

−2000

−1000 0

[Sv]

0 5 10 15 20

[m]

20oN 40oN 60oN 80oN

−5000

−4000

−3000

−2000

−1000 0

[Sv]

0 5 10 15 20

IFEOM meridional overturning circulation (MOC).

0 0.5 1 1.5 [PW] 2

10oN 20oN 30oN 40oN 50oN 60oN 70oN First guess Rif05

IFEOM heat transports.

A full MDT error covariance matrix is also estimated.

Referenzen

ÄHNLICHE DOKUMENTE

This article both adds to some of these debates by arguing for the benefits of using informal and unstructured conversations in qualitative research, reiterating the critical

Author contributions BB has led overall research activities from proposal development to data compilation, data entry and processing, data analysis, and interpretation of the result

In abnormal optimal control problems it is necessary to basically ignore the objective for certain state values in order to be able to determine the optimal control.. In the

In a constraint-based approach to information structure (Vallduv´ı and Engdahl, 1996; Paggio, 2009), the multimodal relation could be represented in terms of structure sharing

However, as discussed in section 4, differences between a mean velocity field derived from drifter data and the model results appear in the representation of the mean flow

For this satellite altime- try referenced to the GRACE geoid is assimilated together with a set of oceanographic data into an ocean general circulation model (OG- CM).. The OGCM that

Ecosystem models have been embedded into two general circulation models of the oceans (HADOM3L and OCCAM). Both are Bryan- Cox based level models the most important difference

Reasonable model–data differences should be smaller than the sum of model standard deviation σ model and data standard deviation σ data :. ⇒ α