Search

We construct long-term time series of Greenland and Antarctic ice sheet mass change from satellite gravity measurements. A statistical reconstruction approach is developed based on a principal component analysis (PCA) to combine high-resolution spatial modes from the Gravity Recovery and Climate Experiment (GRACE) mission with the gravity information from conventional satellite tracking data. Uncertainties of this reconstruction are rigorously assessed; they include temporal limitations for short GRACE measurements, spatial limitations for the low-resolution conventional tracking data measurements, and limitations of the estimated statistical relationships between low- and high-degree potential coefficients reflected in the PCA modes. Trends of mass variations in Greenland and Antarctica are assessed against a number of previous studies. The resulting time series for Greenland show a higher rate of mass loss than other methods before 2000, while the Antarctic ice sheet appears heavily influenced by interannual variations.

This investigation has had four main thrusts: 1) The analysis of seasonal variations of the Earth's gravitational field using Lageos 1 and 2 SLR data and comparisons to geophysical models. We have estimated the annual variation of the gravity field via a spherical harmonic expansion complete to degree and order 4. We have also constructed a similar model using models of the annual variation in the gravity field due to atmospheric, hydrologic, and ocean mass redistribution. These three models, when combined together, are in excellent agreement with the variations observed by satellite laser ranging. An article on these results was published in the journal Geophysical Research Letters. 2) The second thrust of our investigation has been to analyze the output of a Global Climate Model (GCM) to determine if the GRACE gravity mission can be expected to detect climate change signals. Working with Gary Russell at the Goddard Institute for Space Studies (GISS), we have determined that there are several large secular signals that GRACE might be able to detect, including secular changes in snow cover, sea ice, polar ice, ocean mass, and other variables. It is possible that some of these signals could be detected with 5 years of GRACE measurements - its hard to judge this because the interannual variability in the GCM, which could mask the climate signals, is unreliable. Certainly a follow-on GRACE mission could detect these signals when compared to the data from the initial GRACE mission.). An article on these results will be published in the journal Journal of Geophysical Research. 3) In the last year of the investigation, we developed a new technique for analyzing temporal gravity variations using "geophysical fingerprints", which was successfully demonstrated on 20 years of satellite laser ranging data [Nerem et al., 20031. 4]. We also participated in a workshop on future satellite gravity measurements, which resulted in paper on measuring ocean mass variations using GRACE [Nerem et al., 20031 and on using laser interferometry for future gravity missions [Bender et al., 20031].

Analysis of the gravity and topography of Mars presently provides our primary quantitative constraints on the internal structure of Mars. We present an inversion of the long-wavelength (harmonic degree less than or equal to 10) gravity and topography of Mars for lateral variations of mantle temperature and crustal thickness. Our formulation incorporates both viscous mantle flow (which most prior studies have neglected) and isostatically compensated density anomalies in the crust and lithosphere. Our nominal model has a 150-km-thick high-viscosity surface layer over an isoviscous mantle, with a core radius of 1840 km. It predicts lateral temperature variations of up to a few hundred degrees Kelvin relative to the mean mantle temperature, with high temperature under Tharsis and to a lesser extent under Elysium and cool temperatures elsewhere. Surprisingly, the model predicts crustal thinning beneath Tharsis. If correct, this implies that thinning of the crust by mantle shear stresses dominates over thickening of the crust by volcanism. The major impact basins (Hellas, Argyre, Isidis, Chryse, and Utopia) are regions of crustal thinning, as expected. Utopia is also predicted to be a region of hot mantle, which is hard to reconcile with the surface geology. An alternative model for Utopia treats it as a mascon basin. The Utopia gravity anomaly is consistent with the presence of a 1.2 to 1.6 km thick layer of uncompensated basalt, in good agreement with geologic arguments about the amount of volcanic fill in this area. The mantle thermal structure is the dominant contributor to the observed geoid in our inversion. The mantle also dominates the topography at the longest wavelengths, but shorter wavelengths (harmonic degrees greater than or equal to 4) are dominated by the crustal structure. Because of the uncertainty about the appropriate numerical values for some of the model's input parameters, we have examined the sensitivity of the model results to the planetary structural model (core radius and core and mantle densities), the mantle's viscosity stratification, and the mean crustal thickness. The model results are insensitive to the specific thickness or viscosity contrast of the high-viscosity surface layer and to the mean crustal thickness in the range 25 to 100 km. Models with a large core radius or with an upper mantle low-viscosity zone require implausibly large lateral variations in mantle temperature.

Free-air and Bouguer gravity anomalies from a 50x50 field (MGM635), derived at the Goddard Space Flight Center, with global topography, geology, and the distribution of large impact basins was compared. The free-air gravity anomalies were derived from re-analysis of Viking Orbiter and Mariner 9 tracking data and have a spatial resolution of 250-300 km. Bouguer anomalies were calculated using a 50x50 expansion of the current Mars topography and the GSFC degree 50 geoid as the equipotential reference surface. Rotational flattening was removed using a moment of inertia of 0.365 and the corrections from Table B2 of Sleep and Phillips. Crustal density and mean density were assumed to be 2.9 and 3.93 gm/cm(sup 3). The spherical harmonic topography used has zero mean elevation, and differs from the USGS maps by about 2 km. Comparisons with global geology use a simplified map with about 1/3 the number of units on the current maps. For correlation with impact basins, the recent compilation by Schultz and Frey was used.

Free-air and Bouguer gravity anomalies from a 50x50 field, derived from re-analysis of Viking Orbiter and Mariner 9 tracking data and using a 50x50 expansion of the current Mars topography and the GSFC degree 50 geoid as the equipotential reference surface, with the Martian crustal dichotomy are compared. The spherical harmonic topography used has zero mean elevation, and differs from the USGS maps by about 2 km. In this field the dichotomy boundary in eastern Mars lies mostly at -1 to -2 km elevation. Bouguer gravity anomalies are shown on a map of Noachian, Hesperian, and Amazonian age terrains, simplified from current geologic maps. The map is centered at 300 deg W to show the continuity of the dichotomy boundary. Contour interval is 100 mgals. Gravity and topography were compared along approximately 40 profiles oriented parallel to the dichotomy boundary topographic gradient, to determine how the geophysical character of the boundary changes along its length and what this implies for its origin and development.

New spherical harmonic models of the gravity and topography of Mars place important constraints on the structure and dynamics of the interior. The gravity and topography models are significantly phase coherent for harmonic degrees n less than 30 (wavelengths greater than 700 km). Loss of coherence below that wavelength is presumably due to inadequacies of the models, rather than a change in behavior of the planet. The gravity/topography admittance reveals two very different spectral domains: for n greater than 4, a simple Airy compensation model, with mean depth of 100 km, faithfully represents the observed pattern; for degrees 2 and 3, the effective compensation depths are 1400 and 550 km, respectively, strongly arguing for dynamic compensation at those wavelengths. The gravity model has been derived from a reanalysis of the tracking data for Mariner 9 and the Viking Orbiters, The topography model was derived by harmonic analysis of the USGS digital elevation model of Mars. Before comparing gravity and topography for internal structure inferences, we must ensure that both are consistently referenced to a hydrostatic datum. For the gravity, this involves removal of hydrostatic components of the even degree zonal coefficients. For the topography, it involves adding the degree 4 equipotential reference surface, to get spherically referenced values, and then subtracting the full degree 50 equipotential. Variance spectra and phase coherence of orthometric heights and gravity anomalies are addressed.

The effects of the permanent tidal effects of the Sun and Moon with specific applications to satellite altimeter data reduction are reviewed in the context of a consistent definition of geoid undulations. Three situations are applicable not only for altimeter reduction and geoid definition, but also for the second degree zonal harmonic of the geopotential and the equatorial radius. A recommendation is made that sea surface heights and geoid undulations placed on the Topex/Poseidon geophysical data record should be referred to the mean Earth case (i.e., with the permanent effects of the Sun and Moon included). Numerical constants for a number of parameters, including a flattening and geoid geopotential, are included.