Geoscience Reference
In-Depth Information
Fig.
2
), can now be determined with new and unprecedented accuracy around &3cmat
100 km spatial resolution (Bruinsma et al.
2013
). In comparison to the use of the reference
geoid obtained from the Earth Gravitational Model 2008 (EGM2008), this yields a factor 2
improvement in the MDT at this spatial resolution. However, this accuracy is not neces-
sarily applicable to the Arctic Ocean and the neighboring sub-Arctic seas due to the
presence of sea ice, lack of Jason altimeter coverage and shorter dominant spatial scales.
The GOCE high-level processing facility (HPF) delivers the level-2 global gravity
model from which geoid heights can be determined (Koop et al.
2007
; Bingham et al.
2011
). Based on 12 months of GOCE data acquired in the time interval November 01,
2009 to April 14, 2011, three versions of GOCE gravity model are made available: the
direct (DIR) approach; the spacewise (SPW) approach; and the timewise (TW) approach.
More details of these gravity field models can be obtained from Bruinsma et al. (
2010
) and
Pail et al. (
2011
). In addition, so-called combination models such as the EIGEN-6C (F¨rste
et al.
2011
) that combines the GOCE data with terrestrial data have been developed. In this
paper, we apply the EIGEN-6C gravity model for the computation of the MDT. The
corresponding geoid is determined in the mean-tide system and relative to a Topex-
ellipsoid. This ensures consistency with the Technical University of Denmark (TUD) MSS
data set referenced to the time period 1993-2009 (Andersen and Knudsen
2009
). Sub-
sequent to subtracting the geoid from the MSS, filtering was carried out eliminating the
short wavelength geoid signals, in order to obtain a useful estimate of the MDT. This
filtering was carried out using a 80-km Gaussian filter to preserve the upper bound of the
mesoscale features in the study area. (Note that Knudsen et al. (
2011
) applied a 140-km
Gaussian filter to determine the global ocean MDT.) In the forthcoming, we refer to this as
the GOCE-based geoid and MDT.
Isolines of constant MDT (MSS-G) are usually considered as a stream function for the
large-scale ocean surface circulation, which the surface geostrophic currents are directed
along. In the Northern Hemisphere (Southern Hemisphere), the flow is clockwise (anti-
clockwise) around the topographic high. The magnitude of the global spatial MDT vari-
ations is around 2-3 meters, which is about two orders of magnitude smaller than the
global spatial changes in the marine geoid and the MSS. This makes the computation of the
MDT and the handling of errors challenging as it is easy to fail to exploit all of the details
in the geoid and the MSS when calculating the MDT because of the need to obtain a
smooth solution. Herein, the separation of the MDT from the MSS and the geoid is carried
out in the space domain, where the MSS is usually represented using processing tools that
are available at the dedicated ESA GOCE User Toolbox (GUT); see Web site
http://earth.
The GOCE-based MDT shape and spatial pattern representing the mean from 1993 to
2009 for the North Atlantic, Nordic Seas and the Arctic Ocean is shown in Fig.
3
. The total
MDT elevation range from the high in the Arctic Ocean to the low in the subpolar gyre in
the North Atlantic reaches about 0.9 m. The regional shape of the MDT with the orien-
tation of the dominant slopes in the different sub-domains reveals the presence of the main
circulation pathways in: (1) the subpolar gyre south of Greenland; (2) the inflow of Atlantic
Water, respectively, between Iceland and the Faroe Islands and between the Faroe and
Shetland Islands; (3) the continuous northward flowing Atlantic Water toward the Arctic
Ocean; (4) the southward flowing East Greenland Current (EGC); (5) the Beaufort Gyre;
and (6) the transpolar drift in the Arctic Ocean.
The MDT in the Arctic Ocean may display some characteristic features that are caused
by problems in the data coverage. Both the GOCE data and the altimeter data do not cover
the Arctic Ocean entirely, so within 300-400 km from the pole, the data coverage is
