Assimilation of Earth Rotation Parameters to determine ocean mass change
Jan Saynisch, Manfred Wenzel and Jens Schr ¨oter
Jan.Saynisch@awi.de Alfred Wegener Institute for Polar- and Marine Research, Bremerhaven, Germany GODAE-OSTST Meeting, Nov. 10-12th, 2008, Nice
Abstract
Earth rotation parameters (ERP) are measured with very high accuracy. Changes in the ERP originate in movements of mass within the Earth system. Therefore, these changes can be used to distinguish the eustatic from the steric effects in sea level change. We were able to assimilate measured ERP into a global circulation model of the oceans. On interannual timescales the model shows realistic behaviour and succeeds in the reproduction of the ERP observations. The biggest impact here is on the total ocean mass variation. By simultaneous assimilation of oceanographic data as SSH and SST in addition to ERP conclusions about the ocean heat content could be drawn.
Introduction
Changes in Earth’s rotation rate and axis are described by the Earth rotation parameters (ERP). These changes are induced by mass movements within all of the Earth’s subsystems (e.g. Atmosphere, Ocean, Land). The ERP are measured with very high precision. Non- oceanic influences were filtered out and the remaining signal is translated into ocean angular momentum functions (χ). These were assimilated into a global ocean model to reveal infor- mation about the ocean’s mass changes and movements. TOPEX sea surface height (SSH) measurements constrain the change in the ocean’s volume. A combined ERP and TOPEX assimilation determines the ocean’s heat content anomalies.
I Angular Momentum Functions
−5e−085e−08
Observation without ERP with ERP
X
−1.5e−070.0e+00
Excitation Y
1994 1996 1998 2000 2002 2004
−2e−090e+00
Z
t [years]
Two assimilation runs were made and com- pared to the observations:
• Assimilation of TOPEX, REYNOLDS, LEVITUS
• Assimilation of TOPEX, REYNOLDS, LEVITUS and ERP
➽
II Ocean Mass Anomaly III Ocean Mass distribution (J2)
1994 1996 1998 2000 2002 2004
0e+002e+164e+166e+16 0.000.050.100.15
without ERP with ERP
SSH anomaly [m]
Mass anomaly [kg]
t [years]
1994 1996 1998 2000 2002 2004
0e+004e−108e−10
without ERP with ERP
t [years]
➽
IV Sea level Anomaly(smoothed) V Ocean Heat Content
1994 1996 1998 2000 2002 2004
−0.02−0.010.000.010.02
Observation without ERP with ERP
t [years]
SSH [m]
1994 1996 1998 2000 2002 2004
0e+002e+234e+236e+23 0.000.020.040.060.080.10 warming [K]
without ERP with ERP
t [years]
heat anomaly [J]
➽
➽
➽
➽
ERP Observations
2000 2002 2004 2006 2008
−0.50.00.51.01.5
t [years]
LOD [ms]
Length of Day Anomaly
−200 −100 0 100 200 300
100200300400500
−−> to Greenwhich [mas]
−−> to 90E [mas]
1.1.2000
4.9.2008 3.5.2004
Polar Movement
✚
Model & Method
Ocean model: 2 × 2°(LSG)
Assimilation Method: Adjoint (4D-Var)
ERP Reduction: Highpass filter (2 years)
Atmosphere model subtracted (ERA-40) Land hydrology model subtracted (HDM) Assimilated Data: ERP (IERS)
SSH (TOPEX)
T,S (Reynolds, Levitus)
VI Difference between the two runs VII SSH ERP (χ
z) Correlation
−2e−090e+002e−09χz
1994 1996 1998 2000 2002
−5e+125e+12
t [years]
Freshwater flux[m3 ]Cummulative −1.0
−0.5 0.0 0.5 1.0
0.0 0.2 0.4 0.6 0.8 1.0
0.0 0.2 0.4 0.6 0.8 1.0
longitude latitude
➽
Conclusions
• Monthly to interannual ERP signals can be well reproduced by the ocean model (I)
The assimilation of the ERP observations leads to a negative trend in the ocean’s total mass and distribution (II, III)
Trend and low frequency TOPEX constrains can be well reproduced (VI)
The fit to (I) and (IV) renders the steric contribution to SSH higher than previously modelled (V)
In our model the χ
x,y Signal is mainly generated by the ocean currents (not shown)
In our model the χ
z Signal is mainly generated by surface freshwater flux (VI)
Still, the correlation between SSH and χ
z is close to 1 in the main zonal current systems (VII)