Transcript 13_Deng_OSTST2007 - CSIRO Marine and Atmospheric

Assessment of Geoid Models off Western
Australia Using In-Situ Measurements
X. Deng
School of Engineering, The University of Newcastle, Australia
R. Coleman
School of Geography and Environmental Studies, University of Tasmania, Australia
1
K.R. Ridgway
CSIRO Marine and Atmospheric Research , Australia
Introduction
The Leeuwin Current (LC)
in the Indian Ocean off
Western Australia flows
rapidly poleward.
It is a narrow and
meandering current near
the continental shelf break
from 22°S.
The spatial structure of the
mean flow in the LC is
influential down to the
mesoscale.
A full scale of mean ocean
dynamic topography is
needed in the LC.
2
Introduction (cont.)
To understand the ocean water flowing in the LC,
it is crucial to know the following fundamental
quantities:
–Marine geoid – a gravitational potential surface
determined by the Earth’s gravity field.
–Mean sea surface (MSS) - measured by satellite
altimeter sea surface height measurements.
–Ocean’s mean dynamic topography (MDT) - the
dynamic height difference between the marine geoid
and MSS.
3
A Regional Geoid
A gravimetric geoid model AUSGeoid98 (2'2') exists over
the Leeuwin Current.
– It was computed using data from
• the EGM96 global geopotential model,
• the 1996 release of the Australian gravity database (both
over land and in marine areas),
• a national-wide digital elevation model (DEM), and
• satellite altimeter-derived marine gravity anomalies.
– The geoid height is related to GRS80 ellipsoid.
– Comparisons of AUSGeoid98 with GPS and Australian Height
Datum heights across the continent indicate an RMS agreement of
±0.36 m.
– However, no comparison results presented over oceans.
4
Global Geoids
The recently released Earth geoid models
from GRACE have reached ~20 cm
accuracy at the spatial scale of ~350 km.
–The geoid GGM02C can be used to estimate
the temporal mean circulation at the spatial
scales ~300 Km (Jayne, 2006).
• However, we want to know how it works in
coastal areas.
–EIGEN-GL04S1 (satellite-only and released in
2007)
5
Objectives
To quantify the accuracy of the geoid,
especially the regional geoid in the LC.
To assess the geoid using independent
dynamic heights from oceanic in situ data.
6
Data – Altimeter
Measurements
~9.5 years of TOPEX
SSHs from Jan 1993 to
Jul 2002 (cycles 11 –
364).
–All near coastal data with
the Alt-Bad flag setting
have been deleted.
–22 tracks are selected.
7
Data – Geoid, MDT Models and MSS
Geoid models:
– The AUSGeoid98 (2'X2') model (Featherstone, 1998).
– The GGM02C from GRACE data (to degree and order 200).
MDT models:
– The CSIRO OFAM (1992 – 2004) regional MDT.
– The CSIRO Atlas of Regional Seas (CARS) (Ridgway et al., 2002).
– The Global Ocean Data Assimilation Experiment (GODAE) MDT (Rio
and Hernandez, 2004).
CLS01 MSS model (2'X2').
8
Along-track Comparison among MDTs
An along-track MDT is computed by subtracting the
AUSGeoid98 from altimeter derived along-track mean
profiles.
– Along-track mean sea surface profiles are computed using TOPEX
data from cycles 11 – 364.
– A procedure of the collinear and outlier detection is used for
computing the along-track mean profiles.
The MDT is then compared with existing MDTs.
– The MDT at any along-track point is interpolated using the MDT
grid.
– A filtering procedure was applied to both the MSS and geoid with a
filter radius of 333 km (related to degree and order 60).
9
Along-track Mean Profile Comparisons among MDTs
MSS–AUSGeoid98
10
OFAM
Rio’s MDT
3
2.5
MSS-Geoid
OFAM
Rio's MDT
MDT (m)
2
1.5
1
0.5
0
-31.0 -30.5 -30.0 -29.6 -29.1 -28.6 -28.1 -27.7 -27.2 -26.7 -26.2 -25.7 -25.3 -24.5
Along-track (a101) Latitude (degree)
Comparisons among MDTs along-track a101.
- Biases exist among 3 data sets.
11
- MSS-AUSGeoid98 shows a similar trend to others.
CARS and the MDT Grid Comparison (Mean-Tide System)
CARS
12
MSS CLS01 – AUSGeoid98
CARS and the MDT Grid Comparison (Mean-Tide System)
CARS
13
MSS CLS01 – GGM02C
14
CARS - (CLS01-AUSGeoid98),
CARS - (CLS01-GGM02C),
standard deviation:1.86 m
standard deviation:1.23 m
Conclusions
When using CARS as an independent MDT around Western Australia, the
MDTs computed from the CLS01 MSS and geoids (both regional and global)
do not agree well and show larger standard deviations of 1.86 m and 1.23 m,
respectively.
The MDT computed from the Topex along-track MSS and AUSGeoid98 shows
a larger bias (or systematic error) ~1 m when compared to other MDTs.
The bias also exists between different MDTs.
When using recently geopotential model GGM02C, the omission error caused
by the spatial filtering procedure is observed. Therefore, it is important to find
an effective filter in coastal oceans.
The filtered geoid GGM02C at the spatial scales of ~330 km is still too coarse
to describe the current characteristics of the LC
Further investigation of the accuracy of geoid models over a range of
wavelengths, and improvement of the regional geoid are needed.
15
Thank you and questions
16