Flechtner F.M., Gruber Th., G?ntner A., Mandea M., Rothacher M., Sch?ne T., Wickert J. (Eds.) System Earth via Geodetic-Geophysical Space Techniques

Подождите немного. Документ загружается.

Sea Level Variations – Prospects from the Past

to the Present (SEAVAR)

Tilo Schöne and Jens Schröter

Climate change and climatic ﬂuctuations are a natural phenomenon and have

occurred over different temporal scales during the Earth’s history. In addition to this

natural variability, the question of human-induced changes of climate has gained

increasing public awareness. This interest is caused by indications of recent global

warming, which has been related to increasing portions of anthropogenic green-

house gases in the atmosphere. The importance of sea level when studying climate

change is, that it responds as a highly sensitive indicator of climate change, the main

causal connections being the mass exchange between land ice and ocean water and

the thermal expansion of the ocean water.

Besides this monitoring function of climate change, sea level itself is of public

concern. This is a consequence of the fact that a large portion of the human pop-

ulation lives close to the coast and, therefore, would be directly affected by a sea

level rise as expected for global warming. Thus, irrespective of whether the current

changes are governed by natural or by anthropogenic processes, it is of prime inter-

est to determine sea level and sea level changes with high accuracy, to study the

likely sources and to predict future sea level changes.

Measuring sea level variations with tide gauges and radar altimetry and modeling

the associated effects by general ocean circulation models is the primary means to

derive consistent statements about regional and global sea level change patterns.

Moreover, in combination with the new series of gravity ﬁeld models derived from

CHAMP and GRACE, steps are made towards the separation of mass and density

changes for parts of the world ocean.

For the last two decades altimetry has played an important role in detecting sea

level change on a global scale. A strong regional variability even for decadal peri-

ods has been measured from satellites. A main pitfall of altimetry is that no single

mission covers the whole time span with sufﬁcient accuracy. The series of radar

altimetry satellites started in 1985 with the launch of GEOSAT but was discontinued

in 1989 till 1992. Since than a series of high quality satellite missions have been

T. Schöne (B)

Helmholtz Centre Potsdam, GFZ German Research Centre for Geosciences,

Department 1: Geodesy and Remote Sensing, Telegrafenberg, 14473 Potsdam, Germany

e-mail: t.schoene@gfz-potsdam.de

313

F. Flechtner et al. (eds.), System Earth via Geodetic-Geophysical Space Techniques,

Advanced Technologies in Earth Sciences, DOI 10.1007/978-3-642-10228-8_24,



Springer-Verlag Berlin Heidelberg 2010

314 T. Schöne and J. Schröter

followed or paralleled each other. However, the analysis of the results of all missions

does not lead to unique sea level estimates. Depending on the choice of corrections

algorithms distinctly different trends are found. With the consistent reprocessing

of satellite orbits (chapter “Radar Altimetry Derived Sea Level Anomalies – The

Beneﬁt of New Orbits and Harmonization” by Schöne et al., this issue) for the

ﬁrst time comparable accuracies for sea level anomalies and mean sea level was

achievable.

The derivation of long-term trends is complicated by the fact, that climate-

induced decadal and secular sea level changes can be concealed by seasonal, annual

and interannual variations, which may act as noise masking long-term trends. Long-

term time series of sea level measurements are available only from tide gauges

(TGs), which are also sensitive to vertical land movements and are, thus, records

of relative rather than absolute sea level. Another difﬁculty comes from the inho-

mogeneous distribution of TGs over the globe. An optimal analysis of TG records

requires assimilating the data into a global ocean model with the goal of extrapolat-

ing these data over the entire domain in a dynamically consistent way. However,

most of global ocean circulation models are not capable of resolving the ocean

coastline in a sufﬁciently accurate way which is also necessary to fully exploit the

ﬁne resolution of the Gravity Field and Steady-State Ocean Circulation (GOCE)

Mission.

The accurate determination of sea level, the prediction of its future changes and

its use as an indicator of climate change is linked to our understanding of the gov-

erning processes in the strongly interactive dynamic system Earth. This confronts

scientists with the following tasks:

• Accurate determination of absolute sea level using multi-mission satellite altime-

try supported by space and surface techniques from GEOSAT till today (chapter

“Radar Altimetry Derived Sea Level Anomalies – The Beneﬁt of New Orbits and

Harmonization” by Schöne et al., this issue).

• Estimation of vertical rates of GPS benchmarks at TGs and combination with

TG measurements to derive absolute sea level change records with high resolu-

tion and long history (chapter “Reanalysis of GPS Data at Tide Gauges and the

Combination for the IGS TIGA Pilot Project” by Rudenko et al., this issue).

• Improved estimates of historical global sea level variability by combining TG

measurements and radar altimetry (RA) through empirical orthogonal functions

(EOF) (chapter “A 15-Year Reconstruction of Sea Level Anomalies Using Radar

Altimetry and GPS-Corrected Tide Gauge Data” by Schön et al., this issue).

• Estimation of regional and global sea level change for 20 years and more to ﬁll

the gap between the GEOSAT and TOPEX/Poseidon altimeter missions by ocean

modeling and assimilation (chapter “Sea Level Rise in North Atlantic Derived

from Gap Filled Tide Gauge Stations of the PSMSL Data Set” by Reinhardt et al.,

this issue).

• Reﬁnement of ocean model resolution for better description of boundary currents,

assimilation of absolute dynamic topography in conjunction with hydrographic

measurements from autonomous proﬁling ﬂoats and preparation of the use of

Sea Level Variations – Prospects from the Past to the Present (SEAVAR) 315

a ﬁne resolution geoid (chapter “Using ARGO, GRACE and Altimetry Data to

Assess the Quasi Stationary North Atlantic Circulation” by Richter et al., this

issue).

• Consistent assimilation of altimetry derived sea level anomalies from GEOSAT

to TOPEX/Poseidon into an ocean circulation model (chapter “Combining

GEOSAT and TOPEX/Poseidon Data by Means of Data Assimilation” by Wenzel

and Schröter, this issue).

A major step forward made in SEAVAR is the consistent reprocessing of satel-

lite orbits (chapter “Radar Altimetry Derived Sea Level Anomalies – The Beneﬁt of

New Orbits and Harmonization” by Schöne et al., this issue) starting from GEOSAT

(1985–1989), ERS-1 (1992–1996), ERS-2 (since 1995) to TOPEX/Poseidon

(1992–2005). Small inconsistencies still exist since the older satellites had different

satellite tracking technologies. Nevertheless, the main improvement came from the

consistent use of up-to-date gravity ﬁeld models and improved tracking station coor-

dinates. The new orbits for the ﬁrst time allowed estimating high quality sea level

anomalies for the period covering 1985 till now. This improvement had allowed

to (i) assimilating a long time series into the general ocean circulation model and

(ii) to estimate sea level variability over the past 15 years using a combination of

radar altimetry and tide gauges.

To construct long time series of sea level anomalies for e.g., the past 50 years,

tide gauges can be used as a proxy. PSMSL (UK) provides revised tide gauge

time series for a global set of gauges; however, the time series are affected by

local vertical land movement. In 2001 the Tide Gauge Benchmark Monitoring Pilot

Project (TIGA) was initiated by the International GNSS Service and GLOSS to

process and provide estimates of vertical land movement. GFZ is contributing from

the beginning, providing consistent time series of vertical motion at tide gauges.

Within SEAVAR a new attempt has been made to reprocess a much larger data

set of GPS data at tide gauges (chapter “Reanalysis of GPS Data at Tide Gauges

and the Combination for the IGS TIGA Pilot Project” by Rudenko et al., this

issue). Now, vertical estimates for more than 280 sites are available. In combina-

tion with solutions of other contributors, a global and uniﬁed GPS solution is now

available.

Both the GPS corrected time series and the newly derived sea level anomalies

have been used to construct sea level anomalies (chapter “A 15-Year Reconstruction

of Sea Level Anomalies Using Radar Altimetry and GPS-Corrected Tide Gauge

Data” by Schön et al., this issue). A major contribution is the study of the sensitivity

of the reconstruction to the geographical distribution of the tide gauges, which is

of importance for reconstructions of sea level anomalies over longer periods back

in time.

Combining the altimetry data and tide gauge data with ocean circulation mod-

els opens further perspectives in assessing sea level rise over time spanning

GEOSAT and TOPEX/Poseidon missions. Data assimilation studies carried out at

AWI (chapter “Combining GEOSAT and TOPEX/Poseidon Data by Means of Data

Assimilation” by Wenzel and Schröter, this issue) rely on tide gauge data to bridge

316 T. Schöne and J. Schröter

the gap between the missions and ﬁnd that ocean warming accounts for one third of

the global sea level rise and concludes that the inﬂow of fresh water adds another

2 mm/year.

A reconstruction of the sea level rise for the North Atlantic relying on tide gauge

data for 1950–2006 and the Inverse Finite Element Ocean Model (IFEOM) show

1.18 mm/year sea level rise for this period (chapter “Sea Level Rise in North Atlantic

Derived from Gap Filled Tide Gauge Stations of the PSMSL Data Set” by Reinhardt

et al., this issue). About 27% are explained by steric expansion of the ocean while

the remaining 73% are due to inﬂow of water from land or surrounding oceans.

SEAVAR successfully demonstrated the advantage of a consistent and combined

approach in using radar altimetry, tide gauges, GPS and ocean circulation models.

Radar Altimetry Derived Sea Level

Anomalies – The Beneﬁt of New Orbits

and Harmonization

Tilo Schöne, Saskia Esselborn, Sergei Rudenko, and Jean-Claude Raimondo

1 Introduction

Satellite radar altimetry (RA) has proven to be a reliable and efﬁcient method to

monitor the global sea level and its short- and long-term changes. With currently ﬁve

active RA missions (ERS-2, JASON-1, ENVISAT, GFO-1, JASON-2), the global

sea level can be observed with reasonable accuracy. Exploiting the already com-

pleted missions of GEOSAT, ERS-1 and TOPEX/Poseidon sea level time series of

more than two decades can be constructed. Due to the different orbital parameters,

accuracy, long-term system stability and spatial coverage, all satellites, however,

give slightly different results in terms of regional sea level change and sea surface

variability.

In the past several attempts have been made to calibrate and long-term monitor

individual as well as to cross-calibrate different RA missions. For JASON-1 and

TOPEX/Poseidon this has been successfully demonstrated (e.g., Ablain et al.,

2009). However, the poor orbit quality of especially GEOSAT has heavily con-

strained studies on decadal sea level variability. With the recently developed gravity

ﬁeld models EIGEN-GRACE04Sp (Foerste et al., 2008) and improvements in orbit

modeling new orbits have been computed for various RA missions (GEOSAT,

ERS-1, ERS-2, TOPEX/Poseidon). For GEOSAT the improvement is most clearly

visible, reducing the RMS of single crossover differences from 15 to 9 cm which is

almost the level of ERS-2.

2 The Altimeter Database and Processing System (ADS)

For accurate analyses of local and global sea level it is vital to use consistently

processed data from active and completed altimeter missions incorporating state

of the art correction models. In order to integrate, process, update, and harmonize

T. Schöne (B)

Helmholtz Centre Potsdam, GFZ German Research Centre for Geosciences,

Department 1: Geodesy and Remote Sensing, Telegrafenberg, 14473 Potsdam, Germany

e-mail: t.schoene@gfz-potsdam.de

317

F. Flechtner et al. (eds.), System Earth via Geodetic-Geophysical Space Techniques,

Advanced Technologies in Earth Sciences, DOI 10.1007/978-3-642-10228-8_25,



Springer-Verlag Berlin Heidelberg 2010

318 T. Schöne et al.

data from different radar altimeter missions the Altimeter Database and Processing

System (ADS) has been developed since several years at GFZ. Using the ADS sys-

tem it is possible to correct the data for various instrumental and environmental

effects as well as to merge different correction models optimized for the desired

application.

The ADS consists on the one hand of a database holding data from most com-

pleted and recent RA missions (GEOSAT, ERS-1, ERS-2, TOPEX, GFO-1, Jason-1)

as well as state of the art instrumental, orbital, and environmental correction models.

This structure allows for integrating upcoming data corrections and new correction

models in a timely manner. The parallel storage of different correction models for

the same effects allows for testing their inﬂuence on derived quantities, such as mean

sea levels and crossover statistics.

In addition ADS provides tools to extract and process data – starting from the

extraction of uncorrected data or correction models at given coordinates and times

up to the derivation of statistical quantities like crossover point statistics, collinear

analysis, and the derivation of mean sea surface height and sea level anomaly

models. The data is provided along-track as well as interpolated onto regular geo-

graphical grids. A web interface (http://adsc.gfz-potsdam.de) allows external users

to retrieve and process data and to derive many statistical quantities.

3 Harmonization of Different Altimetric Missions

Almost all past and recent RA missions are using different processing standards

and different environmental correction models. As long as single mission studies

are performed, these differences can be neglected. However, when comparing or

combining different RA data, a harmonization is necessary.

Some environmental effects can be corrected using the same correction mod-

els, e.g. tidal effects or the reaction on atmospheric pressure changes. For those

effects, state of the art correction models are provided consistently for all missions

available at the ADS. Other correction models, like the ionospheric or wet tropo-

spheric correction, rely on the satellite hardware. Harmonization is done by using

the most accurate corrections available for the individual missions. For the selection

of correction models s tudies have been performed to identify systematic differences

between the different models. Other correction models including e.g. sea state bias

and orbits it is possible to re-calculate them at least in a similar manner. In this paper,

however, we will concentrate on the consistent recomputation of satellite orbits and

their effects on sea level anomalies derived from GEOSAT data.

For the calculation of mean sea level and sea level anomaly maps discussed here

the best-available correction models, orbits, and data upgrades (e.g., Rudenko and

Schöne, 2007; Rudenko et al., 2006; Scharroo et al., 2000; Scharroo et al., 2004;

AVISO, 2008) have been used. In case several state of the art models were avail-

able the “best model” was selected by minimization of single- and inter-mission

crossover differences (Figs. 1 and 2). It can also be seen that the sea level trends

Radar Altimetry Derived Sea Level Anomalies 319

1985 1990 1995 2000 2005

–4

–2

Height [m]

GEOSAT

TOPEX

JASON-1

ERS-2

ERS-1

GFO

Fig. 1 Sea level time series

derived using the selection of

best-performing correction

models and reprocessed orbits

1994 1996 1998 2000 2002 2004 2006 2008

–3

–2

–1

height [cm]

Fig. 2 Sea level time series

derived from the combination

of TOPEX and JASON-1

altimetry

still differ between the missions. These are artifacts originating from orbit insufﬁ-

ciencies, problems in t he external calibration, as well as from some of the correction

models.

4 The Effects of New Orbits

Precise orbits of RA satellites GEOSAT (March 1985–December 1989), ERS-1

(August 1991–June 1996), TOPEX/Poseidon (September 1992–October 2005) and

ERS-2 (May 1995–February 2006) were derived using GFZ’s EPOS-OC software

in the same ITRF2000 reference frame (Boucher et al., 2004) for all satellites using

common models (Table 1) for the time periods given above. The orbits were com-

puted using the following types of observations: GEOSAT orbit was calculated

using Doppler and single RA crossover (SXO) data, ERS-1 and ERS-2 orbits using

satellite laser ranging (SLR) and SXO data and TOPEX/Poseidon orbit using SLR

and Doppler Orbitography Integrated by satellite (DORIS) observations.

320 T. Schöne et al.

Table 1 The main models, input data and corrections used for the RA satellite orbit determination

Forces, input data, corrections Model used

Geopotential EIGEN-GRACE04Sp up to n = m = 80 (70 for

TOPEX/Poseidon) (Foerste et al., 2008)

Third body gravitational attraction Sun, Moon, Mercury, Venus, Mars, Jupiter, Saturn

(DE405, LE405) (Standish, 1998)

Atmosphere density MSIS-86 (Hedin, 1987), CLS F10.7 cm solar ﬂux and

Ap coefﬁcients

Solar radiation pressure Umbra, penumbra modeled

Albedoandinfraredradiationof

the Earth

Knocke and Ries (1987)

Solid Earth tides IERS Conventions 2003 (McCarthy and Petit, 2003)

Ocean tides FES2004, up to n = m = 50 (Letellier, 2004)

Atmospheric tides Biancale and Bode (2006)

Earth rotation parameters IERS EOP 05 C04 series with IERS2003 daily and

sub-daily corrections

(http://www.iers.org/MainDisp.csl?pid=36-9)

Nutation model Wahr model with VLBI corrections (McCarthy and

Petit, 2003)

Initial station coordinates ITRF2000, if available, or estimated

Station velocities ITRF2000, if available, or NNR-NUVEL1A model

Troposphere correction Marrini-Murray (1973) with Herring correction for SLR

data, Hopﬁeld (1969) model for Doppler and DORIS

data

Ocean loading effect on stations GFZ, based on FES2004 ocean tide model

Atmospheric loading effect on

stations

GFZ, based on ECMWF atmosphere data

Relativistic corrections Post-Newtonian correction, relativistic corrections for

SLR measurements (McCarthy and Petit, 2003)

The 7-day orbital arcs with 2-day overlaps were used for GEOSAT, ERS-1 and

ERS-2 and 12-day arcs for TOPEX/Poseidon. Shorter arcs were used in case of

orbit maneuvers. The orbit quality was evaluated using root mean square (RMS)

ﬁts of measurements, orbital arc overlaps, RMS and mean crossover differences.

The use of the new models, in particular the EIGEN-GRACE geopotential models

(Flechtner et al., 2006) and an optimal parameterization for each satellite (Rudenko

and Schöne, 2007) allowed improving the satellite orbits signiﬁcantly (Rudenko

et al., 2006, 2007), as compared to the orbits derived in the standard distribution

of altimetry data. As a result, e.g., the mean value of RMS crossover differences

reduces from 11.7 cm for ERS-2 orbit derived using PGM055 geopotential and

other rather old models (Massmann et al., 1997) to 7.1 cm for the new orbit.

Improved parameterization and new models also signiﬁcantly enhance the GEOSAT

orbit, especially for the period of high solar activity (MJD 47,400–47,900) (Fig. 3).

The mean value of RMS crossover differences is 15.4 cm for NASA JGM-3 orbit

(GEOSAT Handbook, 1997) and 9.0 cm for GFZ EIGEN-GRACE04S orbit of

GEOSAT.

Radar Altimetry Derived Sea Level Anomalies 321

Fig. 3 RMS crossover differences for GEOSAT orbits computed at NASA (GEOSAT Handbook,

1997) and GFZ; x-axis in Modiﬁed Julian Days (March 1985–December 1989)

Since the GEOSAT original orbit was based on the JGM-3 gravity ﬁeld model

(GEOSAT Handbook, 1997), the progress in orbit modeling, in particular, use of a

new gravity ﬁeld model and improved parameterization signiﬁcantly enhances the

derived sea level quantities especially for the ﬁnal phase (Figs. 4 and 5). Although

–0.8 –0.4 0.0 0.4 0.8

[m] deviation from CLS01 MSSH

–80

–60

–40

–20

–80

–60

–40

–20

0 42 84 126 168 210 252 294 336

Fig. 4 Sea level anomalies of GEOSAT (May 1988) derived using the original JGM-3 based orbits

322 T. Schöne et al.

–0.8 –0.4 0.0 0.4 0.8

[m] deviation from CLS01 MSSH

–80

–60

–40

–20

–80

–60

–40

–20

0 42 84 126 168 210 252 294 336

Fig. 5 Sea level anomalies of GEOSAT (May 1988) derived using the GFZ reprocessed EIGEN-

GRACE04Sp orbits

the global weighted sea level change value remains almost identical, regional dif-

ferences can be observed. Notable differences occur especially in the Northern

Paciﬁc and Northern Atlantic. In addition the half-monthly sea level anomalies

(SLA) are less noisy and, thus more suitable, for assimilation into ocean circulation

models.

5 Summary and Outlook

As a result of the reprocessing and harmonization of GEOSAT and other RA mis-

sion data, a radar altimetry data set is now available for studies of decadal sea level

variability. Using improved or new models and modeling approaches for instrumen-

tal and environmental corrections has further improved the consistency of derived

sea level anomalies. Especially the orbit recomputation for different RA missions

allows studies in a consistent reference frame for the ﬁrst time.

The GEOSAT mission still suffers from a negative trend in sea level. Already

a few attempts have been made to calibrate this mission using tide gauge data but

leaking for the correct geo-referencing of the tide gauges. Having now a reprocessed

orbit, GPS correction for tide gauges, and upgraded altimetry data will allow for

calibrating GEOSAT RA data.

So far, there is still a biannual ﬂuctuation visible in the GEOSAT derived SLA

values. Most likely data from one ore more individual tracking stations are causing

this effect, although no single station was identiﬁed so far.

Acknowledgments ILRS (Pearlman et al., 2002) and IDS (Tavernier et al., 2006) data were

used. GEOSAT Doppler data were provided by C.K. Shum (Ohio). RA data are provided by ESA