Download Report

GOCINO
GOCE in Ocean Modelling
Contract GOCINO No SSA5-CT-2006-030756
Deliverable Number D1.1 (D2)
Title: “Report on GOCE data and recommendations on how to use GOCE data in
ocean modelling.”
Nature: Report
Dissemination level: Public
Status: Draft
Date: 12 December 2007
DOCUMENT CHANGE LOG
Rev.
Date
Sections
modified
1
2007-12-12
All
Comments
Changed by
Draft submitted to EU
Per Knudsen /
Ole Andersen
Background
The Gravity and Ocean Circulation Experiment – GOCE satellite mission is a new type of
Earth observation satellite that will measure the Earth gravity and geoid with unprecedented
accuracy. Combining the highly accurate GOCE geoid models with satellite altimetric
observations of the sea surface height substantial improvements in the modelling of the
ocean circulation and transport are foreseen (see Figure 1 on the principle).
Figure 1: Sketch showing the relationship between the geoid, the Mean Dynamic Topography (MDT – the mean
value of the Dynamic Topography) and the Mean Sea Surface (MSS – the mean value of the Sea Surface
Height).
The overall aim of the EU FP-5 RTD project “Geoid and Ocean Circulation in the North
Atlantic – GOCINA” was to enhance European capacity in Earth observation technologies by
promoting and developing methods for the joint exploitation of the approved European Space
Agency ENVISAT (Radar Altimeter) and GOCE missions for ocean circulation studies and
associated climate modelling and operational data assimilation.
Up to the expected launch of GOCE in early 2008 the gravimetric geoid is in general not
known with sufficient accuracy to allow full use of the massive sea surface height information
which several satellite altimetry missions have regularly provided since the early 90´ies, in
global analysis of the ocean circulation. However, in a few marine regions in the world
sufficient in-situ information about the Earths gravity field exists to compute a more accurate
geoid. The region covering the Northern North Atlantic and the Nordic seas between
Greenland, Iceland, Norway, and the UK is one of those regions. A major goal of the
GOCINA project was therefore to determine an accurate geoid in this region and, thereby,
create a platform for validation of future GOCE Level 2 data and higher order scientific
products.
Furthermore, the new accurate GOCINA geoid was combined with a new accurate Mean
Sea Surface (MSS) creating the Mean Dynamic Topography (MDT), which provided the
absolute reference surface for ocean circulation and heat transport. A major outcome of
GOCINA was to use the new and accurate geoid for improved analysis of the ocean
circulation. This way the GOCINA project demonstrated the extent to which GOCE data will
improve the measuring and monitoring of ocean transports in this vital region in the future.
The objective of the GOCE User Toolbox Specifications study supported by ESA was to
develop – in close collaboration with ESA's HPF effort – algorithms and input and output
specification for the subsequent generation of a user toolbox that is required by the general
science community for the exploitation of GOCE level 2 and ERS-ENVISAT altimetry. The
primary oceanography variable of interest to be provided by a toolbox is the dynamic
topography resulting from the difference between altimetric measurements and the geoid
model. Altimetric MSSH fields would be auxiliary input data set fields from which a
consistently filtered mean dynamic topography need to be computed by the dedicated GOCE
User Toolbox (GUT).
Objectives
The main objective of GOCINO is to complement existing research and development
activities by advancing the readiness of operational oceanographic centres to incorporate
new Earth observation data in their working methods and, hereby, fill a gap between current
GMES projects in order to integrate and fully exploit the approved European Space Agency
GOCE satellite mission for ocean modelling and operational data assimilation.
GOCINO will support the advance of the capabilities in exploitation of EO data from
forthcoming satellite mission GOCE in pre-operational oceanographic services of GMES
through the following specific networking activities:
•
•
•
•
•
•
•
Dissemination of the scientific results from the EU FP-5 RTD project “Geoid and
Ocean Circulation in the North Atlantic – GOCINA”,
Apply GOCINA products and recommendations to develop strategies for
implementation of GOCE products in operational ocean models together with the
ECMWF, TOPAZ, FOAM, MERCATOR, and MFS operational centres,
Facilitate interaction and communication between the GOCE data processing
consortium and the oceanographic users to transfer knowledge and exchange
experiences and requirements,
Promote the exploitation of GOCE data in the operational centres in the EU FP-6
Integrated Project MERSEA using the assimilation strategies for the ECMWF,
TOPAZ, FOAM, MERCATOR, and MFS systems,
Disseminate and transfer the implementation strategies to the MERSEA project to
follow up and coordinate the further implementation of GOCE data into the marine
component of GMES,
Organisation of conferences and workshops, and
Development and maintenance of dedicated web pages for dissemination of
information, knowledge, and experiences.
With GOCINO the knowledge and expertise build up through the research in the GOCINA
project will be kept together and fully made accessible for operational GOCE data users in
the period up to the release of the first GOCE data.
This document aims at providing a recommendation on how to use GOCE data in ocean
modelling based on the results obtained in the GOCINA project. Furthermore, this document
gives a roadmap for how to use GOCE geoid information in ocean modelling. It contains
specific step by step examples from the ESA supported GOCE User Toolbox Specification
(GUTS) study. for the exploitation of GOCE level-2 and ERS-ENVISAT altimetry data. In
GUTS, the objective was to develop algorithms and input and output specifications for the
subsequent generation of a user toolbox that is required by the general science community
Relevance of GOCE data to Oceanography
Satellite gravity measurements are becoming a very important tool in physical oceanography,
with the success of the GRACE mission and the imminent launch of GOCE. Accordingly, it is
becoming important for oceanographers to understand satellite gravity. This is not as straight
forward as might be thought, since there are a number of subtleties of geodesy associated
with the interpretation of gravity data, and the usual product takes the form of a set of
spherical harmonic coefficients. Oceanographers are generally not used to working with
either of these, so the purpose of this chapter is to describe the basics of the relevant
geodetic issues, with particular reference to GOCE and its measurement system
The primary geodetic quantity of interest to oceanographers is the geoid as this can be used
together with the Mean sea surface to get the Mean Dynamic topography (Figure 1). The
geoid is the level surface which would coincide with sea level if the ocean was in a static
equilibrium. It is the surface relative to which slopes must be calculated to determine
geostrophic currents (with a correction for atmospheric pressure gradients). The geoid can
be determined from space by measuring the earth’s gravity field via its effect on the motion of
satellites and of control masses within those satellites.
The geoid is a “horizontal” or “level” surface, a surface which is everywhere perpendicular to
the local direction of gravity. If there were no waves or currents in the ocean, it is where the
sea surface would eventually settle in equilibrium. Since dynamics in the ocean make it
possible for sea level to depart from the geoid, the actual vertical distance of sea surface
height above the geoid is known as the ocean’s dynamic topography (Hughes et al, 2006).
The purpose of the multi-disciplinary GOCINA project was to develop generic tools for ocean
analysis from a simultaneous analysis of (space based) sea surface height and geoid related
observations. Hereby, the project aims at enhancing the European capacity in Earth
observation using data from the approved European Space Agency missions ENVISAT and
GOCE.
The main objectives were first to improve the separate techniques for the determination of
the geoid, the determination of the mean sea surface (MSS), and the determination of the
mean dynamic topography (MDT). Data sets associated with the three quantities should be
compiled and new gravity data should be collected. The qualities of existing models should
be assessed. New geoid, MSS, and MDT models were computed in an optimum way and
assessment and validation of these were performed.
The role of the improved MDT models on the mass and heat exchange across the
Greenland-Scotland Ridge was when examined and the analysis gave invaluable information
on the ocean role in climate. The project will in particular support the GOCE mission with a
set of specific recommendation for integrating GOCE in ocean circulation studies and an
accurate geoid model for validation purposes.
Details on the data processing and how the individual models were obtained are described
below.
Gravity data and geoid
An airborne survey activity was carried out in June 26-July 18 and Aug 7-9 2003. The
measurements were done in a band from Greenland over Iceland and the Faeroe Islands to
Norway and Scotland. A total of 84 airborne flight-hours were flown. The airborne gravity
measurements were processed and incorporated with the revised marine gravity data.
The geoid model was based on the adjusted gravity. Some altimetric gravity data from the
KMS02 model were patched in areas with larger data voids, i.e. more than 20 km to the
nearest marine/airborne data point. A newly released GRACE geopotential model from JPL
was used for the longer wavelengths of the gravity/geoid field in the geoid modelling.
For the error assessment a standard deviation was assigned to the gravity data based on the
nature (terrestrial, marine, airborne) and the history (collector and processor) of the data. The
surface, marine and airborne gravity data was used in a rigorous least-squares collocation
estimate which also provides error estimates on the geoid.
Altimetry and mean sea surface
A new global mean sea surface has been derived using the best available dataset for the
GOCINA region. In deriving this high resolution MSS grid file for the period 1993-2001 with
associated quality indication grid on 2 km or 1/30° by 1/60° resolution the following scientific
achievements were obtained. The MSS model is the only available MSS based on 9 years of
data (1993-2001) using T/P as reference. All data have been interpolated using least
squares collocation taking into account the varying quality and coverage of the data. Global
sea level change over the 1993-2001 was taken into account in the computation. A new
method has been derived to account for the inter-annual ocean variability (like the major ElNiño event in 1997-98), as well as sea surface trends and pressure effects on the ocean
surface. The MSS is available both with and without correction for the atmospheric pressure
correction applied to the altimeter range (inv. barometer correction). Furthermore, annual
averages of the sea level with respect to the nine years mean sea surface have been
computed, so that mean sea surface from other sources and mean dynamic topographies
covering different periods of time may be inter-compared. The MSS model is denoted
KMS04.
Ocean models and mean dynamic topography
Existing mean dynamic topography (MDT) models from ocean models were collected and
reviewed. It was found that the best currently available products were the CLSv2 and
OCCAM, along with the MDT based on the assimilation results available from the UK Met
Office FOAM system. It was decided to compute a composite MDT model using all the
available ocean models (see Table 1). The models were corrected for the differences in
averaging period using the annual anomalies computed from satellite altimetry as described
above. Also, the high resolution models were smoothed to one by one degree grids. The
Composite MDT was derived as the mean
Time period
Resolution
value in each grid node. Furthermore, at each MDT
1993-1999
1°x1°
node the standard deviation was computed to CLS v1
CLS
v2
1993-1999
1°x1°
represent the error of the mean value. The
ECCO
1992-2001
1°x1°
composite MDT and its errors are shown in
ECMWF
1993-1995
1.4°x1.4°
Figure 2.
Table 1: Important features of the MDTs used
in this study
FOAM
OCCAM v1
OCCAM v2
May02-May03
1993-1995
1993-1995
1/9°x1/9°
0.25°x0.25°
0.25°x0.25°
Figure 2. The composite MDT and its associated errors are shown on the left upper and lower panel
respectively. On the right upper panel is the synthetic MDT obtained by differentiating the mean sea surface
and the geoid shown. Lower right the associated errors based on collocation estimates are shown
Combining data sources
The MDT may be obtained simply by subtracting the geoid from the MSS as described in the
previous section. If a full coverage of both the gravity data and the altimeter data exist in the
region, then this simple approach will probably give a nice result. But, if that is not the case,
then more advanced methods may be needed. In the GOCINA region the coverage of gravity
data is not very good in the Northern and the Southwestern parts of the region. For the geoid
computation described above altimetric anomalies are inserted into data gaps, hereby,
combining the two data sources already at this point. To avoid major errors a so-called
draping technique is applied when merging the two data sources.
Both the rigorous and the advanced iterative combination methods have been tested in the
GOCINA region. In the initial tests results were obtained using gravity data and MSS data
only to derive ocean model free MDTs to be assimilated into ocean models. Furthermore, the
characteristics of the MDT error covariances were studied. That was done at a few locations
within the GOCINA region where error covariances were computed. The results showed that
the correlation lengths of the error covariance function all were similar and close to 0.3
degrees. The result of the iterative combination method was obtained using the second
approach.
In reviewing the MDT models the best currently available products were identified. The intercomparisons of those models and MSS and geoid were tested to assess the models. In the
next step the individual models were improved. For the geoid modelling the ship and airborne
gravity data have been merged and compiled with gravity information from the GRACE
satellite mission. The MSS has been developed and updated using ENVISAT altimetry. For
the MDT a composite model was derived using seven individual models. The assessment
and validation of the models were carried out in several steps, e.g. by comparing with other
solutions and with in-situ data. Comparisons along profiles through the region showed that
the new GOCINA models picks up more important details, e.g. the southward current through
the Faeroe-Shetlands Channel.
The models were further improved by integration and optimising of the three independent
technologies – gravimetry, satellite altimetry, and oceanographic ocean circulation models
from the previous work packages. An integrated MDT model that reached a level of precision
similar to the MSS was produced. New techniques were developed to estimate the errors in
the MDT. The MDT model was used for assimilation studies where transports through the
Greenland-Scotland ridge were estimated and forecasting of the ocean circulation improved.
The results of those experiments are described in other GOCINO documents.
On GOCE geoid determination
The GOCE satellite measures the earth’s gravity field in two ways, by satellite-satellite
tracking (SST) plus accelerometer, and by gradiometry. The former is the more familiar
technique (the same as that used by CHAMP). The acceleration of the satellite is due to a
combination of gravitational forces and body forces (such as atmospheric drag and thruster
forces). Using the onboard accelerometers to determine the acceleration due to body forces,
the GPS tracking of the satellite then constrains the estimation of gravitational accelerations,
permitting the earth’s gravitational field to be determined. This technique is particularly suited
to longer wavelength parts of the gravity field. The second method used by GOCE is
gradiometry, and it is this method which permits the recovery of short wavelength features in
the gravity field. Gradiometry uses a pair of accelerometers to measure the difference in
gravitational acceleration between two nearby points (separated by 0.5m for GOCE). There
are three such pairs in GOCE, arranged along mutually orthogonal axes, resulting in a full
measurement of the three-dimensional gradient of gravity (9 numbers, each representing the
gradient of one component of gravity along one particular direction).
The usual way to represent the gravity field is in terms of spherical harmonic coefficients. In
principle, the calculation of geoid height from these coefficients cannot be performed in a
single step, as it involves calculating the potential at an unknown position. In practice it can
be simplified by a linearization about a known position, the reference ellipsoid, since the total
potential W is known to be close to a constant there. The linearized geoid height N above the
ellipsoid is then given by the Bruns formula
N (θ , λ ) =
T (θ , λ )
γ (θ )
where γ is the gravity taken from the reference earth, θ is the geodetic latitude and λ is the
longitude. For geoid heights of up to 100m this approximation is good to sub-centimeter
accuracy (having found the position of the geoid to this accuracy, it is then possible to use
this more accurate position to evaluate the potential much closer to the true geoid, after
which a further application of the Bruns formula results in accuracy well below millimetric).
GOCE data in ocean models: A Roadmap
GOCE is a very ambitious mission, and many conditions for its success lie at the level of the
processing of its data, which for a large part is going to be new to everyone. This implies that
special and dedicated care of the data processing should be taken, to ensure that the best
Earth’s gravity field model can be delivered to the scientific users. For this reason a
European GOCE Gravity Consortium, EGG-C is established. Its main purpose is to
determine the best possible global model of the Earth’s gravity field from the pre-processed
data of the GOCE mission of ESA, with derived grids of geoid heights, free-air gravity
anomalies, geoid slope, together with their uncertainties mapped from the covariance
information on the model parameters, all this after thorough evaluation of the model quality.
The GOCINA project supported the EGG-C tasks in two distinct cases, namely (1) by
educating and preparing the community in using GOCE data for oceanography including sea
level and climate research as well as operational prediction; and (2) by generating a best
possible regional gravity field and geoid model for the North Atlantic that can be used in
validation of the GOCE products.
The task of educating and preparing the scientific community in using GOCE data for
oceanography and a test of the performances of the schematic approach, was done in
providing two proceeding from the GOCE ESA-ESRIN and GOCINA Luxembourg
workshops. A resume of the proceedings is as follows:
The accurate and high-resolution marine geoid, will in combination with precise satellite
altimetry enable new estimates to be made of the absolute ocean topography. In combination
with in-situ data and ocean models, this will, in turn, provide a high-resolution “window” on
the ocean circulation at depth. Developments for merging the gravity information, which will
be obtained by GOCE and other gravity missions, into ocean models have been addressed.
The GOCINA project examined the latter issue confined to the Northeast Atlantic and
southern region of the Nordic Seas. The Arctic Ocean to the north, the deep North Atlantic
Ocean to the southwest, and the shallow North Sea to the southeast bound the region. The
exchange of water masses across the Scotland-Greenland gap has a profound influence on
the thermohaline circulation leading to a horizontal and vertical density structure unlike any
other ocean regions. The question is then how the mean dynamic topography (MDT) reveals
this characteristics structure. To examine this, the GOCINA project produced three surface
fields for the ocean area under investigation, including the mean sea surface (MSS), the
geoid, and the MDT.
The GOCINA MSS, KMS04 was based on 9 years (1993-2001) of different altimeter data.
The geoid was estimated by solving the integral of the gravity field over the Earths surface
using data from a new established database, consisting of new and earlier airborne surveys
and ship data. Similarly, several MDTs was gathered from existing ocean circulation models.
Although they display similar large scale patterns, clear differences are observed at local
scale such as confined to the Irminger Sea, the Scotland-Greenland gap, and the Norwegian
Sea. However, there are several possible sources for the disagreement: the ocean models
are different; the MDTs represent the model means over different integration period; the
spatial resolutions of the models vary; the forcing fields are different.
The simulations of GOCE impact on the gravity field recovery were done using the full
spectra of the signals and the errors, and a GOCE like geoid at an assumed resolution of
100 km with an accuracy of 1-2 cm was used for a number of simulations. Based on mission
parameters and extensive simulations it has been demonstrated that GOCE will meet those
requirements. An important outcome of the simulations is thus a set of error degree
variances that may be included as errors in other simulations of the GOCE performance.
Computation of a GOCE MDT
The ocean Mean Dynamic Topography (MDT) relevant to oceanographers for a chosen
period is the difference between an altimetric Mean Sea Surface (MSS, computed for the
chosen period) and a geoid model N:
MDT=MSS-N
This apparently very simple equation must be handled with care, as the equation is only
applicable when the Altimetric Mean Sea Surface and the Geoid satisfy the following three
conditions:
1. The two surfaces must have the same spatial content
2. The two surfaces must be given relative to the same reference ellipsoid.
3. The two surfaces must use the same tide system
Once these three points have been taken into account and once both the MSS and the geoid
have been adequately processed, the Mean Dynamic Topography can be computed. By
construction, the spectral content of the MDT is therefore limited by the spectral content of
the geoid model. In the case of GOCE, the corresponding MDT will thus have centimetre
accuracy at a 100 km resolution.
To account for the residual geoid signal having wavelengths shorter than about 100 km a
filtering of the MDT values are required. The filtering required can be carried out spatially or
spectrally.
Using the GOCE User Toolbox (GUT)
The GOCE User Toolbox (GUT) provides a powerful tool for accessing GOCE-data. The
accurate and high-resolution marine geoid derived from GOCE data, will in combination with
precise satellite altimetry enable new estimates to be made of the absolute ocean
topography.
The example below shows the GUT as the most simple level and the input parameter and
options to be set in order to derive a specific Mean Dynamic Topograpy. Here the MSS
CLS01 and the EIGENGL04S geoid are used to compute the ocean Mean Dynamic
Topography at a 400 km resolution using a Gaussian filter in geographical space.
The input and output parameters for the workflow are shown in Figure 3.
Figure 3: Example of a MDT computation
The GOCE User Toolbox will be designed so that it can be used at different levels,
depending on the expertise and the needs of the user. The toolbox is made of 6 workflows.
The use of “workflows” is allowing the computation of geoid/gravity field/MDT in one single
step, with few inputs required.
A more general examples are shown below:
When computing the ocean Mean Dynamic Topography, the MSS and the geoid first have to
be computed in the same “system”. This means that:
1. The two surfaces must be given relative to the same reference ellipsoid.
2. The two surfaces must use the same tide system
3. The two surfaces must have the same spatial content
Both the altimetric Mean Sea Surface heights and geoid heights should be given relative to a
common reference ellipsoid, which corresponds to a theoretical shape of the Earth. Figure 4
shows the height differences between the GRIM and the TOPEX ellipsoids on a global grid.
Figure 4: Height difference between the TOPEX and the GRIM ellipsoids.
Also, geoid heights and Mean Sea Surface heights often differ depending on what tidal
system is considered to deal with the permanent tide effects. In the MEAN TIDE system, the
effects of the permanent tides are included in the definition of the geoid. In the ZERO TIDE
system, the effects of the permanent tides are removed from the gravity field definition. In the
TIDE FREE or NON-TIDAL system, not only the effects of the permanent tides are removed
but the response of the Earth to that absence is also taken into account. Altimetric Mean Sea
Surfaces are usually expressed in the MEAN TIDE system. The GRACE GGM02 geoids
from the CSR are defined relative to the ZERO TIDE system. The GRACE EIGEN geoids
from the GFZ are defined relative to the TIDE FREE system.
Figure 5 shows the difference between the TIDE FREE and the MEAN TIDE reference
systems.
Figure 5: Height difference between the TIDE FREE and the MEAN TIDE reference systems
If the use does not take these these considerations into account, the impact on the resulting
MDT is large: for instance,
In order to assist the user in ensuring, that the two surfaces must have the same spatial
content the GOCE User Toolbox provides the user with the MDT computation techniques
allowing integrating short-scale information from other MDT sources. Those techniques will
be further referenced to as Remove-Restore techniques.
Each workflow is a succession of processes that can also be called independently by the
user.
Furthermore, many single functions may be called successively, providing an even more
complex and flexible processing tool
The general workflow associated with the GOCE User Toolbox is shown in Figure 6.
Figure 6: GUT main workflow
Through the use of this workflow, the user has access to the three main outputs of the
toolbox, namely geodetic fields (geoid height, gravity anomaly, deflections of the vertical), a
Satellite-only Mean Dynamic Topography and a Combined Mean Dynamic Topography. All
products are computed using the default procedures and parameters recommended by the
GUTS expert team. For instance, the MDTS is computed in spectral space using a spatial
filter with a default filter width (that will depend on the GOCE data and is therefore not
defined yet – around 100 km).
All outputs are gridded fields (1/2° resolution, regular). When used with the default input
fields (MSSH and a-priori MDT) provided with the toolbox, the default MDTS and MDTC are
obtained.
References
GOCINA (Geoid and Ocean Circulation in the North Atlantic) – Final Report, Danish National
Space Center, Technical Report Number 5, Copenhagen, 2005, ISBN 9788791694073;
ISBN 87-91694-07-8
GUTS (GOCE User Toolbox Specification) - Work Package 5000: “Summary and Tutorial
Document”, ESA/XGCE-DTEX-EOPS-SW-04-0001
Hughes, C. W., Bingham, R. J., 2006: An oceanographer’s guide to GOCE and the geoid.
Ocean Sci. Discuss., 3, 1543-1568
Tapley, B., Ries, J., Bettadpur, S., Chambers, D., Cheng, M., Condi, F., Gunter, B., Kang, Z.,
Nagel, P., Pastor, R., Pekker, T., Poole, S., Wang, F., J., 2005 : GGM02 - An improved Earth
gravity model from GRACE. Geodesy, 79:467-478, doi:10.1007/s00190-005-0480-z.