Hydrological Sciences-Journal-des Sciences Hydrologiques, 44(3) June 1999 407

Gravity and the hydrosphere: new frontier

JEAN O. DICKEY1, CHARLES R. BENTLEY2,

ROGER BILHAM3, JAMES A. CARTON4, RICHARD J. EANES5,
THOMAS A. HERRING6, WILLIAM M. KAULA7,
GARY S. E. LAGERLOEF8, STUART ROJSTACZER9,
WALTER H. F. SMITH10, HUGO M. VAN DEN DOOL11;
JOHN M. WAHR12, MARIA T. ZUBER13
Members of the US National Research Council/National Academy of Science Committee
on Earth Gravity from Space

Abstract Satellite gravity measurements expected in the next few years will provide
unprecedented views of the Earth's gravity field and, given sufficient duration, its
changes with time. Gravity changes directly reflect changes in the masses of the ocean
(thus allowing the separation of steric (heat induced) and non-steric contributions to
sea-level rise), the Greenland and Antarctic ice sheets, and the water stored in the
continents. Not only can measurements of those changes provide a truly global
integrated view of the Earth, they have, at the same time, sufficient spatial resolution to
aid in the study of individual regions of the Earth. These data should yield information
on water cycling previously unobtainable and be useful to both fundamental studies of
the hydrologie cycle and practical assessments of water availability and distribution.
Together with complementary geophysical data, satellite gravity data represent a new
frontier in studies of the Earth and its fluid envelope.
La gravimétrie et l'hydrosphère: de nouvelles perspectives
Résumé Les mesures gravimétriques satellitaires que l'on peut espérer obtenir dans
quelques années fourniront un vision sans précédent du champ de gravité de la Terre,
ainsi que de ses variations au cours du temps. Les variations du champ de gravité de la
Terre reflètent directement les modifications de la masse des océans (permettant de
distinguer les contributions d'origine thermique à l'élévation du niveau des mers), des
couches de glace du Groenland et de l'Antarctique et de l'eau stockée sur les
continents. Ces mesures peuvent non seulement fournir une vision globale de la Terre,
mais elles ont aussi une résolution spatiale suffisante pour permettre l'étude de régions
particulières. Ces données devraient permettre d'obtenir des informations sur le cycle
hydrologique jusqu'ici inabordables, qui seront utiles tant pour l'étude fondamentale
du cycle de l'eau que pour l'évaluation de la répartition et de la disponibilité de l'eau.

Jet Propulsion Laboratory/California Institute of Technology, Pasadena, California 91109, USA,

jean.dickey@jpl.nasa.gov
- University of Wisconsin-Madison, Madison, Wisconsin 53706, USA, bentley@geology.wisc.edu
CIRES, University of Colorado, Boulder, Colorado 80309, USA, bilham@stripe.colorado.edu
University of Maryland, College Park, Maryland 20742, USA, carton@atmos.umd.edu
University of Texas, Austin, Texas 78722, USA, eanes@csr.utexas.edu
Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA, tah@mtglas.mit.edu
University of California, Los Angeles, California 90024-1567, USA, wkaula@ess.ucla.edu
Earth and Space Research, Seattle, Washington 98102, USA, lagerloef@esr.org
Duke University, Durham, North Carolina 27708-0230, USA, stuart@duke.edu
NOAA Geosciences Laboratory, Silver Spring, Maryland 20910-3281, USA, walter@amos.grdl.noaa.gov
NOAA Climate Prediction Center, Camp Spring, Maryland 20746, USA, wd51hd@sgi45.wwb.noaa.gov
University of Colorado, Boulder, Colorado 80309, USA, wahr@lemond.colorado.edu
Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA, zuber@mit.edu

Open for discussion until 1 December 1999

408 Jean O. Dickey et al.

En liaison avec d'autres données géophysiques, la gravimétrie satellitaire ouvre de

nouvelles perspectives à l'étude de la Terre et de son enveloppe fluide.

INTRODUCTION

The Earth's gravity field provides a record of the mass distribution within the Earth's
system that can be used to understand the evolution and dynamics needed to maintain
that distribution. Processes that affect the distribution of mass in the Earth produce
variations in the Earth's gravitational field on a variety of spatial and temporal scales
(Fig. 1). Highly accurate measurements of the Earth's gravity field made with
appropriate spatial and temporal sampling can thus be used to better understand the
processes that move mass within, on, and above the Earth. The inversion of the gravity
signals to obtain the mass distribution and an understanding of the dynamics that cause
them are not straightforward problems. However, through a combination of spatial and
temporal analyses as well as auxiliary land-based measurements for ground truth,
knowledge of the gravity field and its temporal variations can provide insights into the
processes involved. The reliability of these inferences depends on the accuracy, spatial
and temporal resolution, and duration of the gravity measurements.

ADVANCES IN UNDERSTANDING GLOBAL GRAVITY

In a recently released report of the US National Research Council (NRC, 1997), the
Committee on Earth Gravity from Space explored the scientific questions that could be
addressed with a better determination of the global gravity field, examined generic
mission scenarios developed from approximately a dozen mission concepts, and
deduced the expected advances that would flow from each. Traditionally, the gravity
field has been treated as essentially steady-state, or static, over human lifetimes
because over 99% of the anomalous gravity field (the departure of the field from a
simple ellipsoidal model) is static in historic time. The static field is dominated by
irregularities in the solid Earth caused by convective processes that deform the solid
Earth on time scales of thousands to millions of years. Spaceborne gravity
measurements have already led to dramatic advances in recent years in the
understanding of those processes and their relationships to the structure and dynamics
of the core and mantle, the thermal and mechanical structure of the lithosphère, ocean
circulation, and plate tectonics.
While we expect that static-field studies will yield significant future results, we
envision that the most dramatic advances arising from the next generation of gravity
satellites will arise from the examination of the remaining anomalous gravity field.
Temporal variations, which have time scales ranging from hours to secular changes,
are caused by a variety of phenomena that redistribute mass. Many of these are related
to the hydrosphere; processes range from sub-daily oscillations (e.g. tides) to longer-
term trends (e.g. aquifer depletion). The cryosphere (the part of the Earth's surface that
is perennially frozen) also has seasonal and inter-annual variations, as well as a long-
term secular change. Particularly exciting from a hydrological standpoint is the
 Gravity and the hydrosphere: new frontier 409

Spatial and Temporal Scales of Geoid Variations

Secular/
post-glacialA polar icA g|aciers)
Decadal rebound ' ' '

ocean mass flux

Interannual
Seasonal atmosphere ^
Sub-seasonal hydrology: surface and
ground water, snow, ice

Diurnal solid earth and ocean tides coastal tides^,

Semidiurnal

I I I
10,000 1000 100 10
Wavelength (km)

Static Field Measurement Requirements

10-
ocean floor

0 °
£
o
S 4-

2-

1 -

I -1— T-
10,000 1000 100 10 1
Wavelength (km)
Fig. 1 (a) Geophysical phenomena that cause measurable temporal and spatial
variations in the Earth's gravity field (adapted from Bettadpur & Tapley, 1996); and
(b) summary of requirements for static gravity-measurement accuracy as a function of
wavelength (adapted from NASA, 1987).

potential to study sea-level changes and their causes and changes to groundwater and
surface water (including soil moisture) in continental regions (NRC, 1997; Dickey et
al, 1998).
Satellite gravity data that have accumulated over several decades have been used to
produce complex interpretative models of the Earth's gravity field (Nerem et ah,
1995). Doppler tracking of close satellites (e.g. SPOT) and laser tracking of distant
410 Jean O. Dickey et al.

satellites (e.g. LAGEOS), augmented by satellite radar altimetry and GPS

measurements on TOPEX/POSEIDON, have been the primary data source. The
approval of two new missions based on improved measurement techniques places the
study of the global gravity field at the opening of a new era.
CHAMP, a German-led mission with US involvement scheduled for launch in
1999, will utilize a high-satellite-to-low-satellite tracking scheme to improve the long-
wavelength component of gravity. NASA's Gravity Recovery and Climate Experiment
(GRACE; launch date 2001 (Tapley, 1997; Stanton et al, 1998)) features the first use
of still more accurate low-satellite-to-low-satellite microwave tracking. GRACE'S low
altitude (480-300 km) as well as its long lifetime (planned for five years), will provide
measurements of both the static and temporal components of the gravity field with an
unprecedented spatial and temporal accuracy. High-satellite-to-low-satellite tracking is
a scenario where one satellite is at relatively high altitude and other is positioned at a
low altitude, while low-satellite-to-low-satellite tracking is a scheme where both
satellites are located at low altitude, allowing a more robust resolution. In the
remainder of this article, we will focus on expected advances in time-variable-gravity
measurement that we believe will be of interest to the hydrological community.
Advances in other areas have been highlighted by Dickey et al (1998) and a full
treatment is given in the NRC report (NRC, 1997).

GRAVITY MEASUREMENTS FOR THE HYDROLOGICAL APPLICATIONS

Continental Water Variation Gravity missions can provide estimates of changes in

water storage over spatial scales of several hundred kilometres and larger, which would
be accurate to 10 mm or better (Fig. 2). These would be useful to hydrologists for
connecting hydrological processes at traditional length scales (tens of kilometres and
less) to those at longer scales. The main measurables related to these processes are soil
moisture, groundwater and snow pack. The issue of whether GRACE gravity solutions
can be inverted to solve for those changes in water storage depends on the amplitude of
this hydrological gravity signal, and on whether the hydrological contributions can be
separated from changes in gravity caused by other types of mass variability. The
atmospheric signal is a particular problem because GRACE would be incapable of
distinguishing between a change in water storage in a region and a change in
atmospheric mass integrated vertically above that region. The atmospheric
contributions will be removed from the GRACE results prior to the hydrological
solutions by using the global pressure fields available from meteorological forecast
centres.
To illustrate the capabilities of GRACE for providing hydrological information in
the presence of both GRACE measurement errors and contamination from other
signals, Fig. 3 shows the results of an inversion of simulated GRACE data
(reconstructing mass distribution from the gravity signal) from Wahr et al. (1998). For
this inversion, five years of synthetic, monthly GRACE gravity solutions were
constructed, consisting of a realization of the expected GRACE measurement errors,
plus the estimated gravity contributions from soil moisture (using a global, gridded soil
moisture data set constructed as described by Huang et al, 1996), océanographie
 Gravity and the hydrosphere: new frontier 411

1.000 r x * * * * * * * JK)W)I<T(( ww

0.100
AAAAAA1
O0°00o
•g- 0.010 °oooo000«o
Ooooooooooo"
SST errors for 400-km altitude
Z 0.001 L-
s 10 20 30 40
>
Ï 1.000 ^XXXXXx^
a>
***x**xxx X***Xx„
* 0.100 r AAAAAAAAAAAAAAAAAAAAA.„ +
-R
"S i ô S 0 0
°*0 0
A A
^ ^ ^ - A A A A A A A ;
8 0.010
c
I 0.001 SGGE errors for 400-km altitude

10 20 30 40
3
+ Oceans
flO.000 x Continental Water Storage
CO o Antarctica (1 cm, water equivalent)
A Greenland (8 cm, water equivalent)
1.000 x*x*x*x
*xxx
*xx»
X

0.100 L- AAAAAAAAÀÀAAASAIA11 A !! ++++++ '

iAAA;
ooOoOO^o0
0.010 rSoooo,
'°Oooooooo.ooo;

SSI errors for 400-km attitude

0.001
10 20 30 40
Spherical Harmonic Degree
Fig. 2 The degree amplitudes (in mass), expressed as the thickness of a water layer, for
the annually-varying terms (averaged over 5 years of data) from continental hydrology,
oceanography, and changes in Antarctic and Greenland ice mass. To estimate the
Antarctic and Greenland contributions, annually-varying changes in thickness of 1 and
8 cm, respectively, were assumed for the two ice sheets, which are in reasonable
agreement with the results of Bromwich et al. (1993, 1995). (SST refers to a low-
satellite-to-low-satellite tracking mission which is a GRACE-like mission utilizing
GPS; SSI describes results that would be obtained by the low-satellite-to-low-satellite
tracking mission using laser interferometry; correspondingly SSGE refers to results
expected from an extended mission of spaceborne gravity gradiometer.)

processes (using output from the Los Alamos POP ocean general circulation model of
Dukowicz & Smith, 1994), and errors in the atmospheric pressure corrections
(estimated as the difference between European Center for Medium-Range Weather
Forecasts (ECMWF) and National Centers for Environmental Prediction (NCEP)
412 Jean O. Dickey et al.

•oil moisture data

GRACE recovery

iofdata(cm)>t 6.52
I of dfflarenca (cm) « 0.20 (•)

1 1 1 1 1 . , J J • . »
7
hydrology signal (rms) /
100.00 r
/
= errors In GRACE recovery (rms) ' 1
/ ;
/
/
10.00 r
/
t
/
t

S 1.00 r
•

0.10 r 1

:
(ci
0.01
1000 100
averaging radius (km)
Fig. 3 The results of simulations in which synthetic GRACE data are used to recover
the hydrological signal at Manaus, Brazil (Amazon River basin) (from Wahr et al.,
1998). (a) and (b) show five years of monthly values for two averaging radii (the solid
line is the hydrology signal that went into the simulation, and the dotted line is the
signal inferred from the synthetic GRACE data); and (c) shows the rms of five years of
monthly data as a function of averaging radius (the solid line is the estimate from the
hydrology data; the dashed line represents the accuracy of the GRACE results,
estimated as the rms of the difference between the GRACE recovered values and the
hydrology signal.). The simulated geoid data include the GRACE geoid errors, as well
as contributions from hydrology and oceanography, and our estimated errors in the
atmospheric pressure data.

pressure fields, divided by sqrt(2)). The synthetic gravity solutions were used to infer
monthly changes in water storage averaged over a disc centred on Manaus, Brazil (in
the Amazon River basin), and the results are compared in Fig. 3 with the correct
results, i.e. with the results predicted from the soil moisture data set used in the
simulation. The results are shown in terms of the thickness of a water layer. The results
in the upper two panels show comparisons for five years of monthly values for two
different averaging discs (radii of 500 and 230 km). The bottom panel shows the rms
 Gravity and the hydrosphere: new frontier 413

both of the predicted signal at this location and of the errors in the GRACE recovery of
this signal, as a function of averaging radius. Note that the GRACE errors are smaller
than the signal at radii of about 200 km and larger, are smaller than 1 cm at radii in
excess of about 280 km, and reduce to about 2-3 mm at radii of 400 km and larger.
The 2-3 mm limit comes from the estimated atmospheric pressure errors, and is equi-
valent to a 0.2-0.3 mbar rms pressure error. The degradation of the results with decrea-
sing radius is due to the increase in GRACE measurement error with decreasing scale.
The results shown in Fig. 3 are fairly typical of the expected GRACE performance
at most locations. The inferences are that GRACE should be able to provide useful
estimates of changes in continental water storage at scales of a few hundred km and
larger and at time scales of two weeks and longer, and that GRACE has the potential of
delivering monthly values of water mass with accuracies of better than 10 mm in water
thickness at large scales. These water storage estimates can be used both for
monitoring purposes and to improve understanding of hydrological and climate-related
processes. Because GRACE measurements will consist of averages over hundreds of
km, they will complement traditional ground-based hydrological measurements, which
tend to be restricted to the scales of individual catchments. The GRACE results can
also complement space-based microwave and infrared soil moisture measurements.
The latter provide information on only the upper few cm of water in the soil, whereas
the GRACE estimates reflect changes in water storage throughout the water column.
What GRACE does allow is closure of the total water budget over large regions. This
should permit the assessment and improvement of regional-scale hydrological models
and of the hydrological components of climate and meteorological models. It could
prove useful also for helping to monitor the water available for agricultural purposes
and to help in assessing flood and drought danger over large regions.
Sea Level Rise and Glaciology Satellite observations of time-dependent gravity
could constrain changes in global sea-level in two ways. First, they could be used to
help identify the thermal-expansion component of sea-level rise by using the
observations to solve directly for the mass increase in the ocean. From the GRACE
mission (five-year duration assumed), an increasing mass of water in the ocean
equivalent to 0.1 mm year"1 of sea-level rise can be measured. That will be a major
improvement over the uncertainty at present, which is almost an order of magnitude
larger. Second, they could help identify the continental sources of water transferring
into the ocean, through continual monitoring of the mass of ice sheets, continental
glaciers, and groundwater aquifers. Using GRACE to monitor liquid water on land was
outlined above. Probably even more important to sea level are temporal variations in
the masses of glaciers, ice caps, and ice sheets (the term "ice sheet" is reserved for the
vast ice covers of Antarctica and Greenland).

GRAVITY FIELD AND THE CRYOSPHERE

Ice bodies on land can shrink in two ways—by excess melting and liquid water runoff
and by breaking off as icebergs. In either case the mass lost soon reaches the ocean.
Conversely, ice bodies can gain mass only by an excess of snowfall on their surfaces;
the moisture source is evaporation from the oceans. Thus, the measurement of the total
414 Jean O. Dickey et al.

mass balance (deviation from constant mass through time) of the ice on land is a direct,
though partial measure of the changes of water mass in the oceans.
There are two immense advantages of gravity measurements for large-scale mass-
balance studies. The first is that since gravity depends simply on mass, changes in
gravity provide a direct measure of ice-mass balance that is independent of changes in
the mean density of the ice bodies, which can change with time. Secondly, gravity
serves as an automatic integrating tool, thus obviating the huge difficulty that
glaciologists face in extrapolating results from field studies in a few areas to much
larger regions, especially to the entirety of the vast ice sheets.
Changes in the masses of the Antarctic and Greenland ice sheets may be major
contributors to sea-level change. Unfortunately, the large-scale mass balances of the ice
sheets are currently only poorly known. Estimates of their contributions to the global
sea-level rise during this century are in the range of -1.4 to 1.4 mm year"1 for
Antarctica and -0.4 to 0.4 mm year"1 for Greenland (Warrick et al., 1996)—not even
the sign is known.
GRACE will be sensitive to changes in the overall mass of Greenland and
Antarctica to an almost unbelievable accuracy, equivalent to better than 0.01 mm year"1
of sea-level rise corresponding to each ice sheet! However, there are ambiguities in
interpreting gravity signals over ice sheets because of problems in separating the
effects of ice sheet changes from the effects of isostatic rebound (i.e. creep in the
mantle in response to the geologically recent removal of ice sheets) and inter-annual
variations in snow accumulation rates and atmospheric pressure, which preclude the
attainment of that high accuracy. Fortunately, these ambiguities will be much reduced
by the determinations of the present-day changes in the heights of the ice sheets that
laser-altimeter measurements from NASA's planned ICES AT mission will yield. That
is because, for a given mass increase, the ratio of surface-height change to mass change
differs widely for isostatic uplift, changes in solid ice, changes in snow accumulation at
the surface, and atmospheric pressure changes. The combined measurements will make
it possible to determine the contribution of the ice sheets to sea-level rise with an error
of only one or two tenths of a millimetre per year. (For a more extensive discussion see
Bentley & Wahr, 1998.)
Satellite gravity measurements are capable of yielding valuable information also
about the mass balance of individual drainage systems within the Antarctic ice sheet, as
well as of the ice sheet as a whole. Glaciologists could use such information to test
models of ice dynamics, which are essential to the prediction of future sea-level
change. Satellite gravity could be used to study secular, inter-annual, and seasonal
changes in the mass of ice and snow in regions characterized by a large number of
glaciers and ice caps. A prime example is the glacier system that runs from the Kenai
Peninsula in southern Alaska down to the coastal ranges of the Yukon and British
Columbia.

CONCLUDING REMARKS

Satellite gravity measurements can provide unprecedented views of the Earth's gravity
field and, given sufficient duration, its changes with time. Not only can they provide a
 Gravity and the hydrosphere: new frontier 415

truly global integrated view of the Earth, they have, at the same time, sufficient spatial
resolution to aid in the study of individual regions of the Earth. In many cases, the
gravity data will be exploited best by assimilating them, together with other
geophysical data, into geophysical models of known processes. Examples of processes
that require ancillary data to better interpret the gravity data are structure and evolution
of the mantle and plumes, structure and evolution of the crust and lithosphère
(including regional deformation of the sea floor and continents, structure of passive
margins, and the continent-ocean transition), and ice-sheet mass balance.

REFERENCES
Bentley, C. R. & Wahr, J. M. ( 1998) Satellite gravity and the mass balance of the Antarctic ice sheet. J. Gluciol. 44( 147),
207-213.
Bettadpur, S. V. & Tapley, B. D. (1996) Accuracy assessments of dedicated geopotential mapping missions. Supplement
to EOS, Trans. AGU 77(46), F140.
Bromwich, D. H., Robasky, F. M., Keen, R. A. & Bolzan, J. F. (1993) Modulated variations of precipitation over the
Greenland ice sheet. J. Climate 6(7), 1253-1268.
Bromwich, D. H., Robasky, F. M., Cullather, R. 1. & Van Woert, M. L. (1995) The atmospheric hydrologie cycle over the
Southern Ocean and Antarctica from operational numerical analyses. Mon. Weath. Rev. 123(12), 3518-3538.
Dickey, J. O., Bentley, C. R., Bilham, R., Carton, J. A., Eanes, R. J., Herring, T. A., Kaula, W. M., Lagerloef, G. S. E.,
Rojstaczer, S., Smith, W. H. F., van den Dool, H. M., Wahr, J. M. & Zuber, M. T. (1998) Satellite gravity and the
geosphere: contributions to the solid Earth and its fluid envelope. Eos, Trans. AGU 79(20), 237-243.
Dukowicz, J. K. & Smith, R. D. (1994) Implicit free-surface method for the Bryan-Cox-Semtner ocean model. J. Geophys.
Res. 99(C4), 7991-8014.
Huang, J., van den Dool, H. M. & Georgakakos, K. G. (1996) Analysis of model-calculated soil moisture over the US
(1931-1993) and applications to long range temperature forecasts. J. Climate 9, 1350-1362.
NASA ( 1987) Geophysical and geodetic requirements for global gravity field measurements, 1987-2000. Report of a
Gravity Workshop, Colorado Springs, Colorado, USA. NASA Geodynamics Branch, Division of Earth Science and
Applications, Washington, DC.
NRC (1997) Satellite Gravity and the Geosphere: Contributions to the Study of the Solid Earth and its Fluid Earth.
National Academy Press, Washington, DC.
Nerem, R. S., Jekeli, C. & Kaula, W. M. (1995) Gravity field determination and characteristics: retrospective and
prospective. J. Geophys. Res. 100(B8), 15,053-15,074.
Stanton, R., Bettadpur, S., Dunn, C , Renner, K.-P. & Watkins, M. (1998) Gravity Recovery and Climate Experiment
(GRACE) science & mission requirements. Document 327-200, ./PL D-I5928. Jet Propulsion Lab., California
Institute of Technology, Pasadena, California, USA.
Tapley, B. (1997) The Gravity Recovery and Climate Experiment (GRACE). EOS, Trans. AGU 78(46), Supplement, 163.
Wahr, J., Molenaar, M. & Bryan, F. (1998) Time-variability of the Earth's gravity field: hydrological and oceanic effects
and their possible detection using GRACE. J. Geophys. Res. 103, 30205-30230.
Warrick, R. A., Le Provost, C , Meier, M., Oerlemans, J. & Woodworm, P. L. (1996) Changes in sea level. In: Climate
Change 1995: The Science of Climate Change (ed. by J. i \ Houghton, L. G. Meira Filho,
B. A. Callander, N. Harris, A. Kattenberg & K. Maskell) (Contribution of Working Group I to the Second
Assessment Report of the Intergovernmental Panel on Climate Change), 359-^105. Cambridge University Press,
Cambridge, UK.

Received 1 December 1998; accepted 10 February 1999