One of the least understood properties of the Earth system is the variability in the distribution of water mass within ocean basins and among oceanic and continental reservoirs, including ice, ground water, biomass, and lake storage. A number of complementary efforts within this interdisciplinary investigation are attacking this problem.
An assessment of mass and momentum redistribution within the oceans comes from studies of polar motion excitation and satellite altimetry. Roughly 30-40% of polar motion excitation apparently comes from oceanic regions over a broad range of frequencies. Based upon correlation analysis, this oceanic excitation arises from atmosphere forcing of the oceans, probably by redistributing mass within the ocean basins [Kuehne et al., 1993]. A significant amount appears to come from the southern oceans, which are poorly observed. Further quantification can be expected from three distinct types of observations which are sensitive to mass redistribution: changes in the center of mass from SLR data [Watkins and Eanes, 1993; Eanes, 1995]; changes in the geopotential inferred from satellite orbit residuals [Chao and Eanes, 1995; Cheng et al., 1995] and direct observation of sea surface height from satellite altimetry. This latter quantity is obtained from TOPEX/Poseidon altimeter data, in combination with estimates of corresponding atmospheric forcing from scatterometry, and meteorological data. An important matter is to separate the effects of vertical (pressure) and tangential (wind stress) loads on the oceans. Improved estimates of the the response of the sea surface to pressure loading have been obtained by Hoar and Wilson  as well as others, and future studies with TOPEX/Poseidon and other satellite altimetry data should be able to quantify the wind driven part of the mass redistribution. Also, dynamic ocean modeling studies are aiding in assessments of mass balance [e.g., Ponte, 1995a]. Preliminary studies of wind stress from ERS-1 scatterometry and other data have been completed [Salstein et al., 1994].
Beyond the question of redistribution within the ocean basins lies the more complex issue of water redistribution within the Earth system as a whole. This is a fundamental issue pertinent to this and many other EOS investigations. Water plays a major role in the redistribution of surplus heat from the equator to the poles via warm wind-driven ocean currents and as latent heat transported in the atmospheric circulation. Atmospheric water vapor is the most important gaseous absorber of solar and terrestrial radiation, accounting for nearly half of the natural greenhouse effect. The high heat capacity of water makes the oceans an important regulator of the global climate by providing thermal inertia to the climate system over a range of time scales. The spatial-temporal distribution of water stored over the land surface is now known to be an integral component of the climate system, and more generally, the sensitivity of the global climate to land surface hydrology has been well documented.
Geodetic data, including measurements of global sea level, polar motion, and the time variable geopotential are remarkably sensitive to variations in the distribution of water mass. For example, the entire annual motion of the pole can be explained by a mere 3 cm annual change in water load on the Earth's surface, and long term variations in continental water storage should produce detectable drift of the mean position of the pole. However, these and other space geodetic data measure the sum of all contributions from atmosphere, oceans, and land areas, and the difficulty is to separately identifying contributions from ice sheet, aquifer, and other reservoirs, to understand the affiliated changes in mass fluxes, and to use the power of the geodetic data as constraints on mass flux estimates. Such a comprehensive understanding of the hydrologic cycle is likely only with the help of a fully global coupled ocean-atmosphere-land climate model that possesses a closed hydrologic cycle. Toward that goal, algorithms for estimating sub-grid scale processes have been developed [Famiglietti and Wood, 1994]. Further model development and connection to atmospheric data, particularly the NASA assimilated model data [Schubert et al., 1993] is a task in progress.3.1.5. Satellite Observations of Mass Redistribution Within the Earth System
Two of the necessary observables for our investigation, Earth orientation parameters (EOP) and time-varying gravity, require the long-term operation of a globally distributed network of satellite laser range (SLR) stations that track spherically shaped targets, including Lageos-1, Lageos-2, Starlette, Ajisai, Stella, and others. The Lageos class satellites are pre-EOS sensors and are currently being tracked for EOP determination. Time-varying gravitational observables are being derived by Lageos and some of the lower altitude geodetic satellites, such as Starlette. Our EOS investigation requires the availability of these two observables, along with an accurate terrestrial reference frame (TRF).
The observed EOP excitation time series provide three global constraints on the variations of the angular momentum of the atmosphere and oceans including the moments and products of inertia of the mass distribution and the relative angular momentum carried in the winds and currents. Analogous time series (orbital excitations) obtained from analysis of satellite laser range (SLR) observations provide additional global constraints on mass redistribution in the atmosphere and oceans for each satellite observed. Unlike the EOP excitations, which are influenced by winds and currents, the orbital excitation time series are caused only by the mass redistribution. Together, the EOP excitations and the orbital excitations are a powerful tool for separating the motion effects (winds and currents) and the matter effects (mass redistribution). Computation of valid orbital excitation constraints for applications to global change research requires a well defined terrestrial coordinate system, and good models for the orbital variations caused by geodynamic phenomena like tidal perturbations. For our investigation, tidally driven orbital signals must be removed from the observed orbital excitations before application to EOS goals.
In this subsection, we will discuss our scientific results to date on the determination and interpretation of the terrestrial reference frame used for the orbital analysis and estimation of the time-varying gravity parameters. Results regarding SLR-determination of center of mass variations of the TRF are described, followed by a discussion of the comparison of observed orbital excitations to meteorological and geophysical models.188.8.131.52. Terrestrial Reference System and Geocenter Variations
The dynamical origin used for orbital analysis is located at the center of mass (CM) of the planet (solid Earth-atmosphere-oceans-cryosphere), or geocenter [Watkins and Eanes, 1993; Eanes, 1995]. The equations of motion of the orbiting body provide access to this point through the dependence of the gravitational accelerations on the position of the satellite in this reference frame. The SLR observatories, on the other hand, are attached to the solid Earth and actually move with respect to the planet's center of mass due to the redistribution of mass in the atmosphere and oceans. The SLR observations can therefore be used to monitor the motion of the terrestrial reference frame rigidly attached to the stations with respect to the CM of the planet. Figure 6 shows the three geocenter time series determined by Lageos-1 SLR observations. This effect is equivalent to monitoring the degree 1 spherical harmonics of the gravitational field of the solid Earth, atmosphere, and oceans. The observed long-period variations in the geocenter have annual variations that are about the size predicted by results from OGCM and slightly larger than those observed in the atmospheric data. The observation can provide global constraints on mass redistribution within the Earth system. The contributions of ground water, ice sheets, and oceans to these variations are being investigated. Geocenter variations may also be important in the determination of absolute sea level variations using radar altimeter measurements, and investigation into this topic continues.
The station locations that define the terrestrial reference frame are computed by analyzing a multi-year set of Lageos-1 SLR observations to determine the position and velocity of each station with respect to the CM. The International Earth Rotation Service (IERS) combines the results of individual techniques like SLR with all of the other space geodesy techniques to form the International Terrestrial Reference Frame (ITRF). While SLR is the only technique that can monitor the motion of the ITRF with respect to the CM with accuracy better than 1 cm, the other techniques (VLBI, GPS, DORIS) provide improved global distribution of the stations defining the coordinate system. We have conducted an accuracy evaluation of the coordinates used in the computation of the EOP and orbital excitation time series. Results indicated that the SLR-determined TRF agrees with VLBI-determined TRF at the cm level in station positions and at the mm/yr level in station velocities [Watkins et al., 1994].