GRACE Science Team Meeting :: Session A.1: GRACE Geodesy

Posters

(Improving global mass flux solutions from GRACE through forward modeling and continuous time-correlation)
Terence Sabaka

(A comparison of GRACE, CHAMP and COSMIC radio occultation over Australia up to 2008)
Joseph Awange

(10-day gravity field models from GRACE and LAGEOS data)
Richard Biancale

(Acceleration approach based on robust estimation for recovering the Earth's gravity field from high-low satellite-to-satellite tracking data)
Zhicai Luo

(Hydrological and oceanic effects on polar motion from GRACE and geophysical models)
Shuanggen Jin

(Methods for reducing correlated error and signal leakage with application to Greenland)
J.Y. Guo

(More results on high-resolution analysis of GRACE sensor data)
Jakob Flury

(GRACE Gravity Solutions and Regularization)
Himanshu Save

(Challenges in resolving high-frequency signal in series of overlapping GRACE RL04 solutions)
Jennifer Bonin

(Inter-Comparison of GRACE Data Releases from Four Current Processing Standards)
Stephanie Castle

(GRACE Support of SLR Analysis for LARES and the ITRF)
E. C. Pavlis

Session: A.1 - Analysis Techniques
Title: Improved Methodology for Derivation of Time-Variable Gravity Solutions from GRACE
First Author: Frank Lemoine
Presenter: Frank Lemoine
Co-Authors: D.D. Rowlands, S.B. Luthcke, T.J. Sabaka, J.P. Boy, J.J. McCarthy, D.S. Chinn, S.M. Klosko, T.A. Pennington, C.C. Carabajal

Abstract: Analysis of GRACE data requires the best possible forward modeling in order to extract the desired time-variable gravity signal. This has involved application of ocean-atmosphere dealiasing models in various incarnations by the different GRACE analysis centers. We show that the effect of forward modeling of hydrology on mass flux solutions from GRACE and demonstrate that ocean mass estimates for GRACE are strongly affected using both spherical harmonics and mascons. We also present the latest results of our global mascon solutions. Recent improvements in our processing allow us to estimate a 6.5 year times series on a grid of 10396 equal area (2 deg) mascon cells every 10 days. We describe this time series, which involves a simultaneous solution for more than 2 million mascon parameters that are correlated by use of temporal and spatial constraints. We also present our latest results from our regional mascon solutions in the cryosphere for Greenland, Alaska, and Antarctica).

Session: A.1 - Analysis Techniques
Title: ITG-Grace04: the new GRACE gravity field release computed in Bonn
First Author: Torsten Mayer-Guerr
Presenter: Torsten Mayer-Guerr
Co-Authors: Enrico Kurtenbach, Annette Eicker

Abstract: The next GRACE gravity field release using the complete time series is currently being computed at Bonn University.

This release will consist of at least three parts: the high resolution static model, unconstraint monthly solutions and daily snapshots derived by using a Kalman smoother. The innovations compared to the last release include the use of improved background models and the processing of GPS phase measurements. Furthermore, the daily snapshots are applied as de-aliasing product in the calculation of the monthly and static gravity field models. As will be shown in the presentation, this procedure appears to be a very important step towards reaching the GRACE baseline accuracy.

Session: A.1 - Analysis Techniques
Title: Improved time-variable gravity field solutions after correction of systematic errors in the Level-1B KBR/KBRR data
First Author: Martin Horwath
Presenter: Martin Horwath
Co-Authors: Jean-Michel Lemoine, Richard Biancale, Stephane Bourgogne

Abstract: GRACE relies on the inter-satellite K-band range (KBR) or range rate (KBRR) observations. We detect systematic errors in the Level-1B KBR/KBRR data products. The errors are in the correction that relates the original KBR measurements between the two antenna phase centers to the distance between the two satellites' centers of mass. We find evidence that the geometric information entering this correction is biased. The biases, in combination with actual variations of inter-satellite alignment, lead to KBR and KBRR data errors that vary along the orbit. These systematic errors significantly affect solutions for the time-variable gravity field, most notably in the zonal components. While the precise origin of the biases remains to be identified, we are able to parametrize and correct them. This leads to more accurate solutions on temporal gravity field variations. The overall rms error level of our 10-day solutions is reduced by the order of 20%. For the zonal coefficients the error reduction is around 50%. We take Antarctic intra-annual and inter-annual ice mass changes as a case study to demonstrate how the processing improvement affects geophysical inferences from GRACE.

Session: A.1 - Analysis Techniques
Title: A method for estimating global diurnal ocean tides from monthly GRACE gravity fields
First Author: Shin-Chan Han
Presenter: Shin-Chan Han
Co-Authors: R.D. Ray; S.B. Luthcke

Abstract: The gravitational forces due to diurnal tides along the ascending and descending GRACE orbits are roughly the opposite in phase except the polar cap regions. It is expected that the un-modeled and mis-modeled effects of diurnal ocean tides on the monthly mean gravity fields would be cancelled mostly. To highlight the diurnal ocean tides and mitigate the climate-related signals in the monthly mean GRACE gravity fields, we exploited the difference between two independent monthly mean solutions from the range-rate data along ascending and descending tracks separately. Instead of applying smoothing posteriori, the internal smoothing method was designed on the basis of a spectral power law and implemented at the least-square inversion. The solar and lunisolar diurnal tides such as K1, P1, and S1, excluding the lunar tide O1 due to its aliasing period shorter than a month, presented noticeable variations in the monthly ascending and descending difference solutions. Computing the Doodson tidal arguments evaluated along the actual GRACE orbits, we decomposed and estimated the remaining tidal signals with respect to a priori model GOT4.7. The adjustment in the tidal height was small yet significant yielding the maximum amplitude of 5 cm mostly in the Antarctica ice shelves and 1 - 2 cm in general at several hundreds km of spatial scale.

Session: A.3 - GRACE Geodesy
Title: The variations in the lower degree Earth gravity field from SLR and GRACE data
First Author: Minkang Cheng
Presenter: Minkang Cheng
Co-Authors: John Ries and Byron D. Tapley

Abstract: Monthly estimates of J2 from satellite laser ranging (SLR) are supporting the use of the GRACE products to extract the mass variations by correcting the significant aliasing effects in the GRACE-derived J2 coefficients. A time series of the monthly solutions of the geopotential coefficients, up to degree and order 5, was derived from analysis of SLR observations of multiple satellites during the GRACE mission period. This presentation compares the spectrum of the variations in the lower degree and order geopotential coefficients from SLR and GRACE measurements, and uses the observed signals in the results from the SLR data for an evaluation of the large scale ice mass changes. GRACE and SLR estimates of the C21/S21 coefficients define the position of the Earth's figure axis. The constraints on the Fluid Love number, which relates the mean figure axis and mean pole of the Earth, and on the rotation of the core will be discussed.

Session: A.3 - GRACE Geodesy
Title: Long-Term Variability of Low-Degree Gravitational Changes
First Author: Jianli Chen
Presenter: Jianli Chen
Co-Authors:

Abstract: Earth gravity change is caused by mass redistribution within the Earth system, including air and water redistribution in the atmosphere, ocean, land, and cryosphere, and mass variation of the solid Earth (in the core, mantle, and crust). Gravitational change can be quantified by geodetic measurements and numerical climate models. We examine long-term (interannual or longer time scales) variability of degree-2 gravitational variations, C21, S21, and C20 using 4 different techniques, from the Gravity Recovery and Climate Experiment (GRACE), Earth Orientation Parameters (length of day and polar motion), advanced climate models (including atmospheric, oceanic, and hydrologic models), and satellite laser ranging. GRACE time series of C21, S21, and C20 show significant long-term variability. A cross comparison among independent estimates suggests that much of this long-term variability may represent climate change signatures.

Session: A.1 - Analysis Techniques
Title: The role of GRACE in multi-sensor joint inversions
First Author: Roelof Rietbroek
Presenter: Roelof Rietbroek
Co-Authors: J. Kusche

Abstract: Due to its large data-volume and consistency, GRACE plays a crucial role in multi-sensor inversions. Even those geophysical parameters which cannot be estimated by GRACE alone, such as geocenter motion, will benefit from the addition of GRACE data in least squares inversions.

In this study we show the influence of GRACE on two types of joint estimation problems. The first one uses a combination of GRACE, GPS and modelled ocean bottom pressure to estimate weekly spherical harmonic coefficients of surface loading. We show data results and compare the results with local bottom pressure recorders. The second inversion uses GRACE, altimetry and tide gauges to separate the contribution from melting fingerprints and GIA related signals. The fingerprints are derived by applying the sea level equation to several loading patterns. Here we focus in particular on the formal contributions of the different datasets.

Session: A.1 - Analysis Techniques
Title: Improving global mass flux solutions from GRACE through forward modeling and continuous time-correlation
First Author: Terence Sabaka
Presenter: Scott Luthcke
Co-Authors: D. D. Rowlands, S. B. Luthcke, JP. C. Boy, F. G. Lemoine, J. J. McCarthy, D. S. Chinn and S. M. Klosko

Abstract: We have been experimenting with two methods to improve the estimation of global gravity from GRACE. First we explore the use of extended forward modeling. We show that the use of a hydrology forward model (over land areas) improves the a prori fit to GRACE Level 1 tracking data and has a large effect on ocean mass determination. We also explore the use of spatial and temporal constraint equations which are used at the solution stage. When global gravity estimated as mason parameters it is easy to employ these constraint equations which tend to preserve power at high degrees while smoothing. The use of temporal constraints causes all gravity parameters in our multi-year time series to be correlated and requires the simultaneous estimation of over 2 million gravity parameters. We present some results from our time series which is based on a grid of 10396 2 degree equal area mascons estimated every 10 days.

Session: A.3 - GRACE Geodesy
Title: A comparison of GRACE, CHAMP and COSMIC radio occultation over Australia up to 2008
First Author: Joseph Awange
Presenter: Joseph Awange
Co-Authors: Khandu; J. Wickert;T. Schmidt; M.A. Sharifi; B. Heck

Abstract: Australia has been experiencing drought for much of the last decade, which has jeopardized water resources and caused degradation in the most productive agricultural regions. The severity of this drought is partly a result of increased temperatures, thought to be a consequence of climate change associated with global warming.. The monitoring of global warming at regional scales, such as Australia, is therefore essential not only for the proper management of water resources, but also to provide appropriate information for policies being formulated for managing our response to climate change. The tropopause, the atmospheric layer separating the troposphere and stratosphere, has been shown to offer one means of monitoring global warming due to its sensitivity to temperature change. Radio occultation (RO) techniques that employ signals transmitted by the Global Positioning System (GPS) have emerged as an important tool for measuring global tropopause temperature over the past two decades. To exploit this information for Australia, this study compares 45,091 RO temperature profiles from the CHAMP (CHAllenging Minisatellite Payload), GRACE (Gravity Recovery And Climate Experiment) and COSMIC (Constellation Observing System for Meteorology, Ionosphere, and Climate) satellite missions over Australia up to 2008. 80 months of CHAMP, 23 months of COSMIC and 20 months of GRACE data are analysed. GPS RO temperature profiles from CHAMP, GRACE, and COSMIC satellites are validated using radiosonde observations provided by the Bureau of Meteorology (Australia). RO soundings between 2001 and 2007 that occurred within 3 hours and 100 km of radiosonde are used for the comparisons. The results indicate that the GPS RO soundings provide information about the temperature of the tropopause of a comparable quality as radiosonde observations, with deviations of less than 0.5�1.5K. Compared to CHAMP and COSMIC, the GRACE data density is however insufficient. With an increased number of GRACE occultations, therefore, the monitoring of Australia's tropopause to an adequate degree of accuracy will be significantly enhanced.

Session: A.1 - Analysis Techniques
Title: 10-day gravity field models from GRACE and LAGEOS data
First Author: Richard Biancale
Presenter: Richard Biancale
Co-Authors: J.-M. Lemoine; S. Bruinsma; S. Gratton; M. Horwath; S. Bourgogne

Abstract: In 2008 the CNES/GRGS geodesy team reprocessed all GRACE data available in an improved modelling context, which resulted in a new release of 10-day time variable geoid models, and a new mean field (EIGEN-GRGS_RL02_MEAN-FIELD), computed up to degree 160 and including variable terms up to degree 50 (drift, annual and semi-annual sine and cosine, and Sumatra effect). On the basis of this modelling context, new GRACE data are regularly processed. Daily normal equations are stacked into 10-day packs and combined with SLR data from LAGEOS for a better determination of low degrees, and then solved using inversion constraints (but no filtering). The 10-day gravity fields are now available on the BGI website until April 2009 (http://bgi.cnes.fr:8110/geoid-variations/README.html). They are given in spherical harmonics expansion as well as series of maps of geoid heights and equivalent surface water heights. Dealiasing products are also provided. In this poster we will present the modelling context, the strategies of processing (namely the stabilization process used in order to lessen the meridian artefacts) and the improvements obtained, by comparing the results to our RL01 series. As a related matter, we wish to mention that a new source of improvement has recently been identified by CNES/GRGS, concerning correction of biases observed in GRACE data. This topic will be discussed in a dedicated oral presentation.

Session: A.1 - Analysis Techniques
Title: Acceleration approach based on robust estimation for recovering the Earth's gravity field from high-low satellite-to-satellite tracking data
First Author: Zhicai Luo
Presenter: Zhicai Luo
Co-Authors: Bo ZHONG; Jinsheng NING; Haihong WANG

Abstract: High-low satellite-to satellite tracking (SST-hl) data can be available from CHAMP, GRACE and GOCE missions. For the inversion of orbit data and accelerometer data into a gravity field model, several approaches can be used, such as the orbit integration approach, the energy balance approach and the acceleration approach. The acceleration approach based on robust estimation is to be studied for recovering the global gravity field from kinematic orbits and accelerometer data.

After the non-gravitation obtained from the on-board accelerometer and the conservative accelerations computed from force models are subtracted from the satellite accelerations derived from the kinematic orbits, the observational equations can be constructed based on Newton�s second law of motion. Then the geo-potential coefficients, as well as the accelerometer scale and bias factor parameters, can be estimated simultaneously using the acceleration approach. Usually there are some outliers in the kinematic orbits, so the robust estimation method is employed to control the effects of the outliers. The global gravity field model WHUCHAMP-ACC70K with 70 degree and order, are recovered from the kinematic orbits of CHAMP with a sampling rate of 30 second provided by GNSS Center of Wuhan University, which cover a time period of approximately 101 days from 9 July 2003 to 17 October 2003, and the non-gravitational accelerations from the accelerometer data provided by GFZ Information System and Data Centre (ISDC).

The WHUCHAMP-ACC70K model is compared to the models of EGM96, EIGEN-1S, EIGEN-2 and EIGEN-CHAMP03S, the degree variance and geoidal height results show that the total accuracy of WHUCHAMP-ACC70K is better than the model of EIGEN-1S and EIGEN-2, which demonstrated the validity of this method.

Key words: SST-hl, gravity field recovery, acceleration approach, robust estimation

Acknowledgement: This research was supported by the National Natural Science Foundation of China (No. 40874002), the National 973 Program of China (No. 2007CB714405), the National 863 Program of China (No. 2008AA12Z105), and the New Century Excellent Talents Plan of Ministry of Education, China (No. NCET-07-0635).

Session: B.2.b - Studies of the Earth's crustal structure
Title: Hydrological and oceanic effects on polar motion from GRACE and geophysical models
First Author: Shuanggen Jin
Presenter: Shuanggen Jin
Co-Authors: D. P. Chambers; B. D. Tapley

Abstract: Land water and oceans are major contributors to the polar motion excitations, second only to atmospheric mass movement. However, quantitative assessment of the hydrological and oceanic effects on polar motion remains unclear due to the lack of global observations, e.g. global terrestrial water storage (TWS) and ocean bottom pressure (OBP). In this paper, hydrological and oceanic mass excitations to polar motion are investigated using monthly TWS and OBP derived from the Gravity Recovery and Climate Experiment (GRACE), the European Center for Medium range Weather Forecasting (ECMWF) model and the ECCO (Estimating the Circulation and Climate of the Ocean) model for January 2003 until December 2008. Results show that the GRACE-derived OBP and TWS better explain the geodetic residual polar motion excitations for the Px component at the annual period, while the GRACE OBP and ECMWF hydrological angular momentum agree better with the geodetic residuals for the annual Py excitation. GRACE ocean and hydrology excitations better explain the geodetic residuals for the semi-annual Py excitation. However, the JPL ECCO and ECMWF models better explain the intraseasonal geodetic residual of polar motion excitation in the Px and Py components.

Session: A.1 - Analysis Techniques
Title: Methods for reducing correlated error and signal leakage with application to Greenland
First Author: J.Y. Guo
Presenter: J.Y. Guo
Co-Authors: X.J. Duan; C.K. Shum; W. van der Wal; L. Wang

Abstract: We present refined approaches for removing correlated errors in GRACE L2 data and for reducing Gaussian smoothing induced land-ocean signal leakage of mass changes computed from GRACE L2 data. Our correlated error removal approach is of the Swenson and Wahr (2006) type. Our approach is based on the error pattern of the Stokes coefficients (SCs) in the monthly GRACE geopotential solutions. We keep a portion of the lower degree-order SCs with the smallest errors unchanged, and high-pass filter the rest using a moving window technique, with window width decreasing as the error of the SCs increases. Both the unchanged portion of SCs and the window width conform with the error pattern, and are adjustable with a parameter. As compared to published approaches of this type, our unchanged portion of SCs and window width depend on both degree and order in a more complex way. We also suggest a way to select preferred parameters of the correlated error removal filter using a synthetic model. Our leakage reduction algorithm uses as input the mass change data after appropriately smoothing using a Gaussian filter. Based on the assumption that mass change signal over land is far larger than that over ocean along coasts, our method approximately recovers the smoothed mass change signal over both land and ocean as if a regional Gaussian filter with the same smoothing radius were applied over land and ocean separately, in which no signal leakages appear. The side lobe problem does not appear in our approach. As the assumption that mass change signal over land is far larger than that over ocean is not necessarily true along every fraction of coasts, the leakage reduction method should be used on a regional basis, i.e., along the coasts where the assumption holds. The leakage reduction method could also be applied to any selected region of interest if mass change signal outside the region could be assumed as being removed base on hydrological models. We apply our approaches to study the ice/snow mass changes over Greenland during 2003-2008. Instead of fitting for trend and acceleration of ice/snow melt, we investigate the yearly average ice/snow storage differences between successive years. The errors in the annual averages of the SCs are much smaller than those in the original monthly model. Hence a less strong filter associated with less signal distortion could be used to remove the correlated errors in the SCs. For the same reason, a spatial smoothing Gaussian filter with a smaller radius associated with higher resolution results could be used. Also, the year-to-year ice storage changes contain far more information than the trend and acceleration computed by least squares fit, since abnormally more melt in one year may result in a trend and acceleration considering the short time span of GRACE data.

Session: A.1 - Analysis Techniques
Title: More results on high-resolution analysis of GRACE sensor data
First Author: Jakob Flury
Presenter: Jakob Flury
Co-Authors: Tamara Bandikova, Nadja Peterseim

Abstract: Efforts are ongoing to reduce and understand disturbances in the time series of GRACE K-band ranging and accelerometer sensor time series. The analysis aims at the further reduction of disturbance effects in gravity field determination, and intends to support mitigation of noise sources for next generation missions. We discuss systematics in absolute and relative attitude variation of both GRACE spacecraft and related effects on the geometrical range correction. Further on, modeling results for spikes introduced by magnetic torquer rod activation are presented. Updates on other disturbance effects are included.

Session: A.1 - Analysis Techniques
Title: GRACE Gravity Solutions and Regularization
First Author: Himanshu Save
Presenter: Himanshu Save
Co-Authors: Srinivas Bettadpur; Jennifer Bonin

Abstract: This paper discusses the process of the design of the regularization matrix for different flavors of the GRACE estimation problems. The orbit and events during the mission introduce a diversity of data quality and hence the variations in processing techniques. This paper looks at the adaptation of regularization methods to such changes and the regularization of GRACE solutions at several different time resolutions.

Session: A.1 - Analysis Techniques
Title: Challenges in resolving high-frequency signal in series of overlapping GRACE RL04 solutions
First Author: Jennifer Bonin
Presenter: Jennifer Bonin
Co-Authors: Srinivas Bettadpur; Byron Tapley

Abstract: Hydrological and oceanographic analysis with GRACE has repeatedly demonstrated that the results at the annual and interannual level are excellent. However, few people have looked at GRACE results at frequencies faster than that, largely due to the monthly sampling of the released fields. Through the use of an overlapping windowing technique, the frequency resolution of CSR GRACE RL04 has been enhanced, allowing weekly to monthly signals to be analyzed. We compare these results to other high-frequency GRACE series and external data. We determine that while GRACE retrieves appropriate hydrological signal at frequencies lower than 3 cycles/year, the signal-to-noise ratio makes it difficult in most areas to trust GRACE results at frequencies much above that. GRACE does seem able to retrieve valid signal up to 6-10 cycles/year, but only in the few regions where the non-seasonal signal is large compared to the expected error. Particular attention should be paid to GRACE signals with monthly or shorter periods, to make certain what GRACE reports has physical relevance.

Session: B.5.a - Theme: 2nd Hydrology Workshop
Title: Inter-Comparison of GRACE Data Releases from Four Current Processing Standards
First Author: Stephanie Castle
Presenter: Stephanie Castle
Co-Authors: J.S. Famiglietti; C. deLinage;M. Lo; J. Reager

Abstract: GRACE water storage data is currently being used extensively in ocean and land applications from regional to global scales. However, GRACE gravity field solutions from the four main processing centers, GFZ, JPL, CSR, and GRGS, have differences in their processing methods and resulting discrepancies in their final products. End product users must then either 1) arbitrarily favor one product over another, or 2) use some combination of multiple products and estimate the processing influence. Here, we attempt to address this issue by comparing the four data products and quantifying the discrepancies. We compare the GRACE data sets through four descriptive metrics: total ocean & land mass time series; terrestrial storage trend maps; terrestrial storage annual amplitude maps; and selected individual basin case studies.

Session: A.1 - Analysis Techniques
Title: GRACE Support of SLR Analysis for LARES and the ITRF
First Author: E. C. Pavlis
Presenter: E. C. Pavlis
Co-Authors: I. Ciufolini, G. Sindoni, P. M. Hinkey

Abstract: GRACE monthly fields are used to generate reasonably accurate and robust estimates of the annual, semi-annual and seasonal signals in temporal gravity variations for the low degree harmonics of the field. These are in turn used to reduce with higher accuracy the precise Satellite Laser Ranging (SLR) data from geodetic targets like the LAGEOS satellites for the definition of the International Terrestrial Reference Frame (ITRF) and for testing relativistic predictions such as the Lense-Thirring (L-T, "frame-dragging") effect predicted by General Relativity. Accuracy requirements for the ITRF are becoming increasingly more stringent, especially with regards to its origin definition and its scale stability. SLR contributes unique information on the origin, and along with VLBI, for its absolute scale. The precise monitoring of the geocenter and its variations over increasingly shorter intervals is a priority goal in order to meet the requirements of the Global Geodetic Observing System (GGOS) of 1 mm for the definition of the TRF origin at epoch and the 0.1 mm/y limit in its temporal evolution. Such stringent requirements are dictated by the geophysical signals that we are trying to observe with confidence, such as sea level variations. The ILRS reanalysis of the 1983 to 2008 data from the two LAGEOS satellites, contributes to a new definition of the ITRF, ITRF2008, due to be released by the end of the year. The new analysis indicates a much more stable definition of the origin with respect to the geocenter, with a highly diminished secular trend compared to the nearly 2 mm/y for ITRF2005. We will examine what this implies for the MSL rise secular rate estimates. In the near future, ASI's LARES mission is to be launched (in about a year from now), adding one more high quality and well-calibrated geodetic target for TRF development and long-wavelength gravitational signal estimation, in addition to its primary purpose of improving the accuracy of the L-T experiment. We will examine what the current error contributions are and what that means in terms of % error of the L-T parameter estimate. We will also present an updated report on the LARES mission status.

Registration

Housing

UT Sponsored Travel

Sessions & Abstracts

Meeting Program

Meeting Proceedings

Past Meetings