Earth Observation with CHAMP

Recent advances in the spatial and temporal retrieval of land-based cryospheric change information south of 42.5° allow fairly robust construction of forward model predictions of the time-rate of change in gravity. A map-view prediction is presented for the time-rate of change in geoid,

d

N/

dt

that might be retrieved from the currently orbiting gravity space craft, CHAMP (Challenging Mini-Satellite Payload for Geophysical Research and Application) and/or GRACE (Gravity Recovery and Climate Experiment). Complementary computation of the surface gravity change,

dδg/dt

, is also presented. The latter can be recovered from terrestrial absolute gravity measurements. Also, the computed rate of change Stokes coefficients for degree and order

l, m

1–12 may be used as reliable estimates of the Southern Hemisphere cryospheric change contribution to the global low-degree harmonic variability recorded in multidecadal satellite laser ranging (SLR) data sets.

We have used one year of CHAMP data for deriving a gravity field model based on the energy balance approach. In order to avoid the use of any a priori gravity information, purely kinematic orbits have been computed from GPS measurements only. Subsequently velocities have been derived from these kinematic positions by two different methods, namely smoothing splines and Newton- Gregory interpolation. Using the principle of energy conservation, the satellite's positions and velocities are transformed into gravitational potential. CHAMP onboard micro-accelerometry is used to correct for surface forces. For spherical harmonic analysis the so-called direct approach has been implemented using the full normal equation matrix. The model, called TUM2Sp, was found to be a more accurate gravity field than EIGEN-2 model.

This study investigates the effect of temporal gravity field variability on CHAMP and other geodetic satellites. The sensitivity of these satellites to the lower order and degree harmonics is presented along with the dominant tidal periodicities. Lumped harmonics from analyses of CHAMP are discussed with a singular value decomposition identifying the dominant combinations. Temporal variability in lower order and degree harmonics are presented in studies with and without CHAMP. These results are compared against those derived from geophysical data.

GPS-CHAMP satellite-to-satellite and accelerometry data covering 2.5 years of the CHAMP mission period were exploited to generate the global gravity field model EIGEN-3p revealing considerable improvements in both accuracy and resolution with respect to the previous model EIGEN-2. For the year 2001, CHAMP and satellite laser ranging data of four satellites were combined to recover largest scale monthly gravity field variations that are subsequently analyzed for the annually varying constituents. The temporal gravity field variations observed by CHAMP and the SLR satellites are compared in the spectral and spatial domain with geophysically (atmosphere, ocean, hydrology) predicted gravity variations that do not reflect the large observed scattering in the monthly solutions but are of comparable size and distribution on the annual time scale.

The operational Superconducting Gravimeter (SG) network can play an important role for validation of satellite-derived temporal gravity field variations. A comparison shows a quite good agreement between SG and CHAMP results within their estimated error bars. It could be proved that the SG-derived temporal gravity variations are representative for a large area within the µgal accuracy, if the local gravity effects are removed.. The long-periodic tidal waves are well determined by ground measurements, therefore they can be applied as a reference for validation. For further validation, field SG measurements should be carried out in representative areas with large gravity variations (e.g. Amazon area).

The near-polar CHAMP and GRACE satellites are now acquiring vitally important new information on the geoid and gravity field of the polar regions. This investigation demonstrates that CHAMP and GRACE data are dramatically reducing the large gaps in our knowledge of the Arctic region, constraining the long wavelength geopotential (>300 km) and beginning to yield the high accuracy marine geoid which is needed for Arctic oceanographic and sea ice studies. Using a detailed Arctic surface gravity field and an independent altimetric gravity field as benchmarks we have evaluated the intermediate-to-long wavelength (> 300km) component of seven CHAMP and two GRACE satellite-only gravity models such as the GFZ EIGEN, the NASA PGS and UT/CSR. We evaluate, spectrally, the errors in - and differences between - these satellite-only models in the Arctic at wavelengths from 300 to 2500 km. The GRACE models accurately resolve Arctic gravity to

full

wavelengths as short as 500 km while the CHAMP models do so to full wavelengths as short as 1000 km. However the CHAMP models continue to show improved resolution as more and better (e.g. lower elevation) data are incorporated. The best CHAMP models agree well with the detailed Arctic ARC-GP model to an rms (error of commission) of better than 2.06 mGal (gravity)and 31 cm (geoid) for all wavelengths (full) longer than 1100 km. GRACE-only geoids are precise to 40 cm or better (all wavelengths) over large areas of the Arctic. CHAMP and GRACE-based geoids could have the accuracy required to detect (together with altimetry) the poorly known dynamic topography of the Arctic Ocean. As an example, a GRACE/detailedgravity hybrid geoid is presented.

This paper discussed consistency of long wavelength components of gravity field between the EGM96 and CHAMP gravity models and compared the long wavelength components of surface gravity data with ones from the models of CHAMP within the same spatial resolution, based on the 2-D Gaussian low-pass filter in China and its vicinity. The results show that the models of EGM96 and CHAMP are consistent up to about degree 35, while above this degrees the EIGEN-2 may have inferior estimates. The evaluation of the terrestrial gravity data for China and its vicinities by comparisons with the models of CHAMP has confirmed the existence of larger errors and systemic discrepancy.

An extended Jacobi integral describes the energy balance of the motion of a satellite referred to a terrestrial reference frame along its orbit. In addition to its classical form inertia forces and non-conservative force function contributions are included. If force function models and observed satellite orbits are consistent the energy balance sums up to a constant. Deviations from it can be caused either by orbit errors or by insufficiencies in the force function models. Therefore, the energy integral offers itself as a validation tool for consistency checks of force functions and orbit determination results. A basic question is the separation of the various sources of inconsistency. In this paper the theoretical foundation of the validation procedure is presented. It is shown that the validation method can be used to detect deficiencies in the orbit modeling and in the gravity field recovery results. Examples are presented how to separate the various causes of inconsistency. Applications to the results of the CHAMP mission demonstrate the procedure.

The new CHAMP and GRACE global gravity field models provide a significantly improved long wavelength gravity spectrum. These satellite-only models are therefore a good basis for studying long wavelength errors of the terrestrial gravity data, as they can be considered as a completely independent source of information.

In this contribution, the models from the CHAMP and GRACE mission as well as EGM96 are used for the evaluation of a terrestrial gravity data set for Europe. The differences are examined both geographically and spectrally. Different techniques are applied for the evaluation, including spherical harmonic expansions, degree variances and the multiresolution analysis based on spherical wavelets. All techniques confirm the existence of small long wavelength errors in the terrestrial gravity data. The reason for such errors may be various, e.g., lacking or poor quality gravity data in some regions, or effects of datum inconsistencies.

The GeoForschungsZentrum Potsdam (GFZ) operationally provides CHAMP orbit products for various purposes. Here the rapid and ultra-rapid orbits are highlighted. Significant developments in Precise Orbit Determination (POD) for Low Earth Orbiters (LEOs), in particular SAC-C and GRACE besides CHAMP, are described. GFZ also started to generate CHAMP-like rapid orbits for SAC-C with good accuracy. Furtheron improved LEO orbit accuracies are demonstrated by simultaneous orbit solutions of the GPS satellites and one or more LEOs in an integrated approach.

The Rapid Science Orbit (RSO) is introduced into CHAMP gravity field processing as an auxilliary tool in order to facilitate the screening and to establish pre-estimated dynamical parameters, especially accelerometer calibration factors. It is furthermore shown how so-called pseudo-attitude angles help to edit the quaternions delivered by the CHAMP on-board star tracker cameras.

This paper presents a synthesis of the CNES/GRGS routine activities regarding the analysis and calibration of the CHAMP accelerometer data. The impact of instrument operations and out of specification temperature variation on bias and scale factor estimation is demonstrated. Precise dynamic solutions of orbits and calibration parameters have been obtained from the processing of two years of data. The results presented here could be obtained thanks to the high performances of the accelerometer and GPS tracking.

A new method to estimate CHAMP satellite-borne GPS receiver clock bias change is presented in this paper, according to the following approach. The difference between two neighboring epoch phase observations includes real CHAMP-GPS distance change, CHAMP clock receiver bias change, cycle slip and other corrections. Real CHAMP-GPS distance change can be obtained at a relatively high precision from a dynamic orbit. As a result, the CHAMP GPS receiver clock bias and cycle slip can also be determined at a high precision.

Reduced-dynamic orbit determination for spaceborne GPS receivers is a method promising highest accuracy of the estimated LEO trajectories. We compare the performance of different pseudo-stochastic orbit parametrizations (instantaneous velocity changes and piecewise constant accelerations) and probe the range between dynamic and heavily reduced dynamic orbits. Internal indicators like formal accuracies of orbit positions, comparisons with orbits computed at the Technical University of Munich (TUM), and validations with SLR measurements are used to assess the quality of the estimated orbits. For piecewise constant accelerations comparisons between the estimated and the measured accelerations from the STAR accelerometer allow for an additional and independent validation of the estimated orbits.

It is shown by means of an extensive simulation study as well as an experiment using real CHAMP data that it is feasible to accurately estimate non-conservative accelerations from precise GPS-based orbit perturbations. Assuming the availability of high-precision gravity field models, such as anticipated for GRACE and GOCE, an accuracy of better than 50 nm/s

2

seems possible for 30-seconds averaged accelerations. The remaining dominant error sources seem to be GPS receiver carrier-phase noise and GPS ephemeris errors.

The technique of using the evolution of a satellite orbit through resonance to determine the values of appropriate

lumped

geopotential harmonic coefficients has recently been revived, and applied to the triple passage of the Champ orbit through 31:2 resonance. Preliminary results for four pairs of coefficients have been derived rapidly, without using the most precise data (which will be forthcoming). The values obtained are compared with those derivable from various global gravity models (to obtain which, vast amounts of data had to be analysed), and the comparison indicates that the resonance technique remains a competitive one.

We report results on the use of CHAMP observations for Earth's mean and time-varying gravity field solutions and the associated geophysical interpretation. 1.5 years of CHAMP data from May 2001 through February 2003 are used for monthly gravity solutions as well as a mean solution based on 6 months of data, employing the energy conservation principle and an efficient conjugate gradient inversion technique. The mean CHAMP gravity field model, OSU03a, is evaluated using other gravity fields model and various data including GPS leveling and Arctic 5′×5′ gravity anomalies. It is shown that OSU03a agrees well with EIGEN2 (CHAMP) model, and both models exhibit improvement over the polar regions. The estimated 3×3 CHAMP time-varying gravity field model, with the exception of J

3

and 2

nd

order tesseral coefficients, are compared with available solutions using satellite laser ranging (SLR) for the semi-annual and annual components. It is shown that CHAMP temporal gravity solution agrees well with the SLR solutions as well as various geophysical fluid models, including atmosphere, hydrology, and ocean. The correlation coefficients between CHAMP and SLR solutions are 0.6∼0.8, with RMS at 0.7∼0.8 mm.

To obtain an alternative gravity solution to that of EIGEN1S, the author's Singular Value Decomposition(SVD) tool,

P

arallel

LA

rge

S

vd

S

olver (PLASS), was applied to the CHAMP normal matrix

ngl-eigen-1s

{xc[2]} to perform an Eigenvalue Decomposition (EVD) analysis. The EIGEN1S solution is based on the Tikhonov regularization method of approximating the ill-conditioned system of equations in a subspace of lower rank. In the EVD solution, poorly determined linear combinations of parameter corrections are removed in the culpable eigenspace of the unconstrained least-squares normal equation. The selection of eigenvalues to be removed, is based upon a new method and four different common optimization (truncation) criteria. The new method, the Kaula Eigenvalue (KEV) relation, optimizes the removal of eigenvalues to best satisfy Kaula's Rule. The four other techniques are: inspection, relative error, norm-norm minimization, and finding the minimum trace of the mean square error (MSE) matrix. Analysis of the five different EVD gravity fields was performed. Two of them were shown to be comparable to the EIGEN1S CHAMP solution obtained by the GeoForschungsZentrum Potsdam (GFZ) {xc[2]}. The best of the five optimal solutions, that of the KEV, is presented. The number of estimated parameters is 11216.

Temporal variations of the gravity field may act either as a signal or as a source of noise for the current satellite gravity missions. This depends, to some extent, on the parametrization of the gravity field solution. We discuss qualitatively how temporal variations affect satellite gravity products and how their effects may be controlled by an adequate parametrization. We describe a mechanism how unparametrized temporal variations may alias into orbit-parallel spatial patterns of a gravity field solution. While the effect is too small to corrupt static gravity field models like EIGEN-2 or EIGEN-3p it may be a concern for studies on time-variable gravity from consecutive GRACE period solutions. Moreover, time-varying errors in non-gravity parameters such as CHAMP accelerometer corrections may, due to correlations with gravity parameters, cause similar effects as geophysical variations. These issues suggest that an adequate parametrization of the gravity field as a function of space and time needs further study. Eigenvalue analyses of solution normal matrices may be a useful tool for these studies.

The gravity field recovery strategy presented here enables the global recovery of the gravity field combined with a regional focus on geographical areas with rough gravity field features in a consistent way. The global gravity field is modeled by a series of spherical harmonics while the regional gravity field features are represented by space localizing base functions of harmonic spline type. The physical model of the orbit analysis technique is based on Newton's equation of motion, formulated as a boundary value problem in form of an integral equation of Fredholm type. The observation equations are established for short arcs of approximately 30 minutes length. The procedure can be applied either globally or regionally to selected geographical regions. For a regional application the coverage with short arcs should be slightly larger than the recovery region itself to prevent the solution from geographical truncation effects. A proper combination and weighting of the normal equations of every arc combined with a tailored regularization allows a stable solution for the field parameters. This procedure can be adapted to the roughness of the regional gravity field features, the discretization of the gravity field and the sampling rate of the observations. A global gravity field solution ITG-Champ01E has been derived based on kinematic orbits covering 360 days from March 2002 to March 2003. Regional gravity field solution have been determined for selected regions with rugged gravity field features.

The energy balance approach is used for a statistical assessment of CHAMP orbits, data and gravity models. It is known that the quality of GPS-derived orbits varies and that CHAMP accelerometer errors are difficult to model. This makes the selection of orbits for gravity recovery difficult. Here we identify the noise level present in in-situ potential values from the energy balance in an iterative variance-component estimation. This means we solve simultaneously for a spherical harmonic model, for polynomial coefficients absorbing accelerometer drift, and for sub-daily noise variance components. These should be understood in a sense of model consistency. Using dynamic GFZ orbits, results including 92 days in 2002 indicate that for most days the noise is at 0.25–0.3m

2

/s

2

, with notable exceptions. The corresponding gravity model is found close to EIGEN-2, after two iterations. With TUM kinematic orbits and Lagrange-interpolated velocities or TUM reduced-dynamic orbits, we found for preselected data the consistency at the 0.7–0.8m

2

/s

2

(KIN), 0.3–0.35m

2

/s

2

(RD) level; gravity models improve significantly on EGM96. Generally, iterative re-weighting improves the solutions significantly, and ‘trackiness’ is considerably reduced.

CHAMP orbits and accelerometer data are used to recover the long- to medium-wavelength features of the Earth's gravitational potential. In this study we are concerned with analyzing preprocessed data in a framework of multiscale recovery of the Earth's gravitational potential, allowing both global and regional solutions. The energy conservation approach has been used to convert orbits and accelerometer data into in-situ potential. Our modelling is spacewise, based on (1) non-bandlimited least square adjustment splines to take into account the true (non-spherical) shape of the trajectory (2) harmonic regularization wavelets for solving the underlying inverse problem of downward continuation. Furthermore we can show that by adapting regularization parameters to specific regions local solutions can improve considerably on global ones. We apply this concept to kinematic CHAMP orbits, and, for test purposes, to dynamic orbits. Finally we compare our recovered model to the EGM96 model, and the GFZ models EIGEN-2 and EIGEN-GRACE01s.

Spherical wavelets have been developed by the Geomathematics Group Kaiserslautern for several years and have been successfully applied to georelevant problems. Wavelets can be considered as consecutive band-pass filters and allow local approximations. The wavelet transform can also be applied to spherical harmonic models of the Earth's gravitational field like the most up-to-date EIGEN-1S, EIGEN-2, EIGEN-GRACE01S, UCPH2002_0.5, and the well-known EGM96. Thereby, wavelet coefficients arise and these shall be made available to other interested groups. These wavelet coefficients allow the reconstruction of the wavelet approximations. Different types of wavelets are considered: bandlimited wavelets (here: Shannon and Cubic Polynomial (CuP)) as well as non-bandlimited ones (in our case: Abel-Poisson). For these types wavelet coefficients are computed and wavelet variances are given. The data format of the wavelet coefficients is also included.

We compare selected techniques for recovering the global gravity field from precisely determined kinematic CHAMP orbits. The first method derives the second derivatives by use of an interpolation polynomial. The second procedure is based on Newton's equation of motion, formulated and solved as a boundary value problem in time equivalent to a corresponding integral equation of Fredholm type. It is applied to short arcs of the CHAMP orbits. The third method is based on the energy balance principle. We implement the analysis of in-situ potential differences following Jekeli's formulation. The normal equations from the three approaches are solved using Tikhonov-type regularization, where the regularization parameter is computed according to a variance component estimation procedure. The results are compared with the recent satellite-only model EIGEN2 and the first GRACE model GGM01s. All methods provide solutions of the gravity field which represent significant improvements with respect to the reference model EGM96 below degree 50. The quality of the solutions differs only slightly.

Using an efficient and robust combination of a kinematic and reduced-dynamic orbit determination procedure CHAMP GPS data spanning about eleven months are processed and different aspects are addressed. Kinematic solutions are generated with and without elevation-dependent weighting of the observations in order to study the impact on the solution. GPS clock corrections with a sampling rate of 30 seconds and of 5 minutes, both interpolated to 10 seconds, are used. The orbit results are compared with the Post-processed Science Orbits from the GeoForschungsZentrum Potsdam, Germany, with orbit solutions from the Technical University of Munich, Germany, and they are validated with SLR measurements.

Analysis of CHAMP GPS data through a reduced dynamic methodology has yielded precise Cartesian positioning for subsequent analyses for gravity field and geodynamic parameters. Results are accordingly presented of (i) a gravity field recovery, (ii) the bias and scale factors for the three axes accelerometer data and (iii) the apparent thrust forces estimated as parameters over a 13-month period. As a consequence we are able to consider the stability of the accelerometer parameters recovered daily or monthly. Analysis of the thruster data will quantify the magnitude of the additional forces involved and the stability, if any, of these forces over an extended period.

This paper presents an overview of the preprocessing of the CHAMP accelerometer measurements as carried out at GFZ Potsdam. The data are smoothed and the sample rate is changed from 1 to 10 seconds. Additionally, the accelerometer data are combined with attitude measurements and some housekeeping data. Two corrections for the linear acceleration are modelled: The so-called X3 correction and a model for the Lorentz force on the proof mass. The preprocessed data are available as a level-2 product at the CHAMP information system and data center at GFZ Potsdam.

The behaviour of the space-borne CHAMP GPS receiver clock over nearly three years is inspected by the results of the Rapid Science Orbit (RSO) determination and the on-board navigation solution (NAV). The analysis completes and enhances the results of the clock characterization study carried out earlier on a limited number of data. It is shown, that CHAMP on-board events like software uploads and hardware reboots frequently lead to large clock offsets, which are reduced on-board within a few hours. Comparisons of RSO and NAV solutions show, that the RSO clock offset estimates are additionally affected by bias and drift of the reference clocks chosen in the processing. The two independent solutions, RSO and NAV, indicate, that the CHAMP clock is free of a drift in the long term. In the short term, the clock parameters show a standard deviation of 0.4 µ/s and a small negative bias of 0.3 to 0.4 µ/s.

The postglacial rebound (PGR) in Fennoscandia is an attractive signal for the detection of time-variable gravity with CHAMP and GRACE. However, the separation from other geophysical signals is a challenge. We study the influence of oceanographic variations of the Baltic Sea. Using a precise high-resolution oceanographic model we analyze its water mass variations and the resulting geoid signal. The geoid signal is similar to the postglacial rebound signal in spatial pattern and in magnitude. Hence we show that the PGR. detection from satellite gravity data will be corrupted by Baltic Sea oceanographic variations unless they are accounted for. Our study suggests to correct for the Baltic Sea signal using external data.

Here we wish to discuss the improvements obtainable in the measurement of the general relativistic Lense—Thirring effect with the LAGEOS and LAGEOS II satellites, in terms of reliability of the evaluation of the systematic error and reduction of its magnitude, due to the new CHAMP and GRACE Earth gravity models.

We investigate North American crustal structure and mass loads from spectral correlation analysis of topographic, CHAMP and terrestrial gravity data. We use free-air and terrain gravity correlations to isolate tectonically driven vertical motions and mass imbalances of the crust and lithosphere. Specifically, we apply correlation filters to decompose the free-air gravity anomalies into terrain-correlated and terrain-decorrelated components to yield compensated terrain gravity effects that we evaluate for crustal thickness variations. Our results compare quite favourably with the seismically inferred global crustal thickness model Crust5.1 and a 3.4 km rms difference with LITH5.0 over North America. Terrain-correlated anomalies reveal mass excesses and deficits that are interpreted as uncompensated elements of the crust. For Hudson Bay, the average terrain-correlated free-air anomaly suggests that the crustal topography is depressed by about 400 m. Because glacial isostatic adjustment considerations can only marginally account for the depression, we speculate that it may reflect other effects such as a preglacial impact.

The long-wavelength part of non-isostatic topography is supposed to be generated by mantle dynamics and is up to now not well studied. In order to separate the dynamic part from the residual topography (that part of the Earth's topography which is not explained by the crustal model providing isostatic compensation) a correlation analysis is performed between the residual topography and long-wavelength isostatic gravity anomalies. It is found that the correlation between these quantities is positive for spatial wavelengths larger than 4000 km. The resulting correlated part of residual topography is regarded to represent the dynamic topography. Its amplitude was estimated to range from −0.4 to 0.5 km and is on the lower limit of what was estimated from a direct modelling of mantle convection. The calculated dynamic topography may be used as a strong constraint in further numerical simulations of mantle dynamics.

In physical oceanography the slope of the sea surface can be used to calculate a surface geostrophic velocity as a reference for the general circulation and its associated transports. Here we calculate a new mean sea surface (MSS) by combining oceanography, altimetry and gravity data. A first guess MSS is calculated from a dynamical ocean model into which measurements of temperature, salinity, property fluxes and tide gauges have been assimilated. The large scales of this surface are subsequently improved by adding information from satellite altimetry and a ’satellite only’ geoid model like the CHAMP derived EIGEN-1S. The combination takes into account the different error structures of the three sources of data. Simultaneous we estimate large scale corrections for the respectively used geoid models. Comparing the corrections of different geoid models makes it possible to verify the improvement of ’satellite only’ gravity models by the CHAMP mission.

Ocean general circulation models which are constrained by altimetry data usually assimilate only temporal sea-surface height anomalies. It is known that this is not enough to correct the mean ocean state. Here we present first results of assimilating the full (absolute) dynamical topography into a steady state version of a finite element ocean model (FEOM) for the North Atlantic. This makes it possible to notably reduce the model-data misfit especially in the western part of the basin and in the southern Labrador Sea.

We have examined contemporary changes in the geoid about Greenland that result from glacial-isostatic adjustment. These may be divided into contributions from ice-load changes that occurred outside of Greenland following the Last Glacial Maximum and changes in the Greenland Ice Sheet (GIS). The GIS's contribution may itself be divided into past and current parts. For past ice-load changes, the resulting geoid displacement is more dependent upon the recent history of the GIS than on the earth model used. Considering an estimated accuracy for the GRACE temporal geoid signal, regional variability in the present-day mass balance of the GIS may be resolved. This variability significantly affects the geoid power spectrum, giving a signal that may be detected by measurements from gravity space missions more easily than has been proposed by other authors.

The long wavelength geoid height undulations are the result of density variations inside the Earth and the dynamic response of the viscous mantle due to the buoyancy forces resulting in dynamic topography, plus the contribution of isostatic topography due to crustal and lithospheric structure. The dynamic topography is a function of the effective stress transmission inside the earth and is linked to the viscosity of the mantle. We solve the equation of motion for a viscous Earth's mantle assuming an incompressible 6-layer shell model and determine the dynamic response function for geoid, dynamic topography, and (poloidal) surface velocity. The internal density distribution can be estimated from seismic tomography, but since density variations might be of thermal and chemical nature, the s-wave velocity to density conversion factor,

R

ρ\ν

s

, varies throughout the mantle. Based on CHAMP gravity field coefficients and four new seismic s-wave tomography models we search for ranges of radial profiles of viscosity and

Rρ\νSubscript>s

, resulting in a correlation to better than 0.85 to the long wavelength hydrostatic geoid, and to a fit better than 0.6 for gravity, dynamic topography and (poloidal) plate velocity. For purely thermal origin

Rρ\νs

should be between 0.2 and 0.4. Successful models however, show a small or even negative value for

Rρ\νSubscript>s

, between 100 and 300 km depth and a low value of ∼ .1 between 700 and 1200 km, but is otherwise roughly constant with values of ∼ .28. The viscosity is slightly reduced in the asthenosphere and even stronger decreased in the mantle transition zone between 410 and 670 km. Resolution for both, viscosity and conversion factor, is poor below the transition zone down to ∼ 1500 km, but well confined in deeper parts of the mantle, where a viscosity between 30 to 40 1021 Pa s and a conversion factor of 0.28 to 0.32 is found.

Regional geoid undulations are determined from CHAMP data using various locally supported basis functions to assess their respective efficiency, accuracy and multi-resolution representation properties. These functions include (biharmonic) B-spline tensor wavelets (with or without compression), multiquadrics (with or without flexible centering and predetermined smoothing) and radially symmetric truncated polynomials.

It is concluded that the B-spline wavelet model is the computationally most efficient approach. The non-periodic variation of the B-spline wavelets allows one to handle data on a bounded domain with small edge effects, and the piecewise linear version allows one to model the geoid using a patch-wise approach. The use of multiquadrics without centering in the data points and predetermined smoothing constant allows handling of heterogeneously distributed data using global optimization. The linear multiquadrics model fits the data best when comparing the residuals of different models with a fixed number of unknowns. For an efficient data synthesis the nonlinear models are best suited due to their far smaller number of basis functions. The smoothest surface was obtained using the nonlinear polynomial approach, whereas the multiquadrics show peaks and the wavelet models show horizontal and vertical edges in their representations. The linear B-spline wavelets are biharmonic, and the approach is capable of an efficient multi-resolution representation of regional gravity field models combining satellite (CHAMP, GRACE, GOCE) and in-situ data.

Some of the main plasma characteristics are reviewed that a LEO satellite with high orbital inclination encounters during its travel across the terrestrial ionosphere of mid- and low-latitudes. It is the region of highest plasma density in the near-Earth environment. Its properties are predominantly ruled by the geomagnetic field. It will be shown how different ionospheric layers — first of all the E- and F-layer — contribute in different ways to the electrodynamic and thermodynamic behaviour of the highly interacting, complex system comprising the ionosphere, thermosphere, and plasmasphere. The physical description of its phenomena and data interpretation have nowadays to rely to a substantial part on numerical methods and models. New observational methods and space missions have essentially contributed to the recent progress in this field. The CHAMP mission takes part in this progress just as much as the IMAGE, TIMED, and other satellite projects as well as ground-based observation programs. The paper summarises recent developments in ionospheric studies as, e.g., the plasma transport at mid- and low-latitudes, the regular Sq-dynamo and the contribution of the F — region dynamo, the interhemispheric coupling by current systems and plasma flows, pulsations, the equatorial electrojet and the plasma fountain effect, the Appleton anomaly, the near-equatorial plasma bubbles, and further open issues.

Crustal field models from CHAMP magnetic measurements are increasingly stable and reliable. In particular, they now allow for quantitative geological studies of crustal structure and composition. Here, we use a forward modeling technique to infer deep crustal structure of continental regions overlain by younger sediments. For this, a Geographical Information System (GIS) based technique has been developed to model the various geological units of the continental crust. Starting from geologic and tectonic maps of the world and considering the known rock types of each region, an average magnetic susceptibility value is assigned to every geological unit. Next, a vertically integrated susceptibility (VIS) is computed for each unit, taking into account the seismic crustal thickness, as given by models 3SMAC and CRUST2.1. From this preliminary VIS model, an initial vertical field anomaly map is computed at a satellite altitude of 400 km and compared with the corresponding CHAMP vertical field anomaly map. We demonstrate that significant geological inferences can be made from the agreement and the discrepancies between the predicted and observed anomaly maps. In particular, the lateral extent of Precambrian provinces under Phanerozoic cover is revealed.

The magnetic crustal thickness of Greenland and the surrounding area is determined by inversion of gridded values of the magnetic radial component as given by the IDEMM model, which is based on CHAMP and Ørsted data alone, and by the Comprehensive Model (CM4), which is based on satellite and observatory data.

After correcting for the remanent magnetization, we determine the vertically integrated magnetization of the crust. Making some simplifying assumptions about the susceptibility, the thickness of the magnetic crust is determined by iteratively improving an initial crustal thickness model using the equivalent source magnetic dipole method.

Regional magnetic signals of the crust are strongly masked by the core field and its secular variation components and hence difficult to isolate in the satellite measurements. In particular, the un-modeled effects of the strong auroral external fields and the complicated behavior of the core field near the geomagnetic poles conspire to greatly reduce the crustal magnetic signal-to-noise ratio in the polar regions relative to the rest of the Earth. We can, however, use spectral correlation theory to filter the static lithospheric anomalies and core field components from the dynamic external field effects. To help isolate regional lithospheric from core field components, the correlations between CHAMP magnetic anomalies and the pseudo-magnetic effects inferred from gravity-derived crustal thickness variations can also be exploited. Employing these procedures, we processed the CHAMP magnetic observations for an improved magnetic anomaly map of the Antarctic crust. Relative to the much higher altitude Ørsted and noisier Magsat observations, the CHAMP magnetic anomalies at 400 km altitude reveal new details on the effects of intracrustal magnetic features and crustal thickness variations of the Antarctic.

A Global Magnetic Petrology Database (MPDB) is now being compiled at NASA/Goddard Space Flight Center and consists of many thousands of records. The prototype database is located at http://core2.gsfc.nasa.gov/research/terr_mag/php/MPDB/frames.html. The purpose of this database is to provide the geomagnetic community with a comprehensive and user-friendly method of accessing magnetic petrology data via Internet for a more realistic interpretation of satellite (as well as aeromagnetic and ground) magnetic anomalies. MPDB is focused on lower crustal and upper mantle rocks and includes data on mantle xenoliths, serpentinized ultramafic rocks, granulites, iron quartzites and rocks from Archean-Proterozoic metamorphic sequences from all around the world. The definition of crustal magnetic anomalies is improving due to the mini-constellation of three satellites — Oersted, Champ, and SAC-C. Recent lithospheric field models (CM4, MF1, MF2) reveal magnetic anomalies with better resolution, for example in the areas of Iceland, Polar Urals Mountains, and Anabar Shield where we have an excellent magnetic petrology records.

A stratospheric balloon flight at 30 km altitude measured the geomagnetic field intensity along a 6000 km track extending from Kamchatka to near the Ural Mountains. When the CM model was used to remove the main and external fields from the observed data, magnetic anomalies of several 100 nT amplitude and 250 to 750 km wavelength are observed. In the eastern part of the track these anomalies appear to be due to the bodies of up to 5 km depth and magnetizations of 0.12 SI (0.01 cgs).

To determine the effect of crustal thickness variation on satellite-altitude geopotential anomalies we compared two regions of Europe with vastly different values, South and Central Finland and the Pannonian Basin. Crustal thickness exceeds 44 km in Finland and is less than 26 km in the Pannonian Basin. Heat-flow data indicate that the crust of the Pannonian Basin has a value nearly three times that of the Finnish Svecofennian Province. A positive CHAMP gravity anomaly (∼4 mGal) is quasi-coincidental with the CHAMP magnetic anomaly across the Pannonian Basin. CHAMP gravity data indicates a minimum of 3 mGal in southwest Finland. CHAMP magnetic data reveal elongated semicircular negative anomalies for both regions with South-Central Finland having larger amplitude (<−6 nT) than that over the Pannonian Basin, Hungary (<−5 nT). In this latter region subducted oceanic lithosphere has been proposed as the anomalous body. In the former the central part of the negative gravity anomaly covers the northern part of the Baltic Sea basin and Gulf of Finland and underlying two rapakivi provinces plus it coincides with an area of lower crustal thickness. The magnetic anomaly directly correlates with the crustal thickness and inversely with the heat flow and, hence, may be caused either by variation of concentration of magnetite or by the elevated Curie-isotherm of magnetite in the lower crust — upper mantle region.

CHAMP is recording state-of-the-art magnetic and gravity field observations at altitudes raging over roughly 300–550 km. However, the non-uniqueness of the process and satellite anomaly errors severely limit anomaly continuation. Indeed, our numerical anomaly simulations from satellite to airborne altitudes show that effective downward continuations of the CHAMP data are restricted to within approximately 50 km of the observation altitudes while upward continuations can be effective over a somewhat larger altitude range. The great unreliability of downward continuation requires that the satellite geopotential observations must be analyzed at satellite altitudes if the anomaly details are to be exploited most fully. Given current anomaly error levels, multi-field inversion of satellite and near-surface anomalies is the best approach for implementing satellite geopotential observations for subsurface studies. We demonstrate the power of this approach using a crustal model obtained by the inversion of combined near-surface and satellite magnetic anomalies for Maud Rise, Antarctica, in the southwestern Indian Ocean. Our modeling, which also includes regional gravity constraints, suggests that crustal thickness variations and remanent magnetization of the normal polarity Cretaceous Quiet Zone produce the dominant satellite altitude magnehtic anomalies.

Models of the main and external field play a key role in the analysis and interpretation of satellite, airborne, marine and ground magnetic data. Here, we introduce a series of main field models with several new features: (1) External fields are parametrised in SM and GSM coordinate systems, accounting for the geometry of the ring current, the magnetosphere and the solar wind. (2) We use vector data globally, instead of the usual approach of using vector data at low latitudes and scalar data at high latitudes. (3) The angles between CHAMP's star camera and its vector magnetometer are co-estimated in a joint inversion with Ørsted data. (4) The model includes 2nd time derivatives of the magnetic field to account for the non-negligible secular acceleration in the current period of new magnetic satellite data. As inferred from the degree spectrum of the current version POMME-1.4, the secular variation is stable to degree 11 and the secular acceleration to degree 6.

A better parameterisation of the field generated by the large-scale magnetospheric current systems is required to avoid the need for further filtering of the data when modelling the crustal magnetic field. Using Ørsted and observatory data for year 2001, we investigate new parameterisations where internal and external degree one Gauss coefficients are computed daily. These models are compared with models where the external field is parameterised with a linear dependence on the Dst index. For both types of model the overall fits to the data are similar but the rapid variations in the magnetospheric currents and their induced counterparts are better captured using the new parameterisations.

New global magnetic data - Gauss coefficients up to degree and order 5, monthly values from 1980 to 2000, fitted to global data and partly based on high-quality satellite vector data of Magsat and CHAMP/ØRSTED - are processed with a recent non-harmonic downward continuation method (Ballani et al. 2002). Using a weakly conducting mantle and the highly conducting fluid in the outer core we investigate the temporal structure of the 1991 jerk below some geomagnetic stations calculating the component dY/dt at the core-mantle boundary and underneath in different depths of the fluid outer core assuming fluid velocity there. The jerk structure dissolves and differs considerably in magnitude and in phase from the harmonically downward continued component.

Champ total field measurements have been used to develop the new version of the Antarctic geomagnetic Reference Model (ARM). The model was conceived as a tool to evaluate the main field in Antarctica, facilitating the merging of different magnetic surveys carried on in the region from 1960 onwards. Spherical cap harmonic analysis was used to produce the model. Together with data coming from POGO, Magsat, and Ørsted satellite missions, a suitable selection of Champ data based on different criteria was performed to minimise the effect of external fields. The comparison of ARM and other global models with regard to real data demonstrates the validity of our regional model, specially for the representation of the secular variation of the geomagnetic field. Since Champ satellite tracks cover the Geographical South Pole better than other satellite missions, this fact contributed to improve the model in the central region of the cap.

Two methods of utilizing data from magnetic satellites for the secular variation modelling are proposed. Both methods are based on the parametric expansion in time, where natural orthogonal components are used as parameters. As data for such expansion are used time series of coefficients of spherical harmonic models, developed from satellite data for each day.

New satellite data have led to increased resolution models of the geomagnetic secular variation (SV). Simple plots of the spectra of these models suggest an unphysical source depth above the core-mantle boundary (CMB). By taking a ratio of the main field and SV spectra, we argue that this result comes from the chosen spectral definition. The models are consistent with a CMB source for SV.

Using an induction method developed by Martinec et al., (2003) which was refined by Martinec and McCreadie (submitted) to include satellite data, calculations of the magnetic effect of induced currents at CHAMP satellite heights for a region over the Pacific Ocean are shown. The method generally reproduces the data for the northern hemisphere but underestimates the inductive effect in the southern hemisphere. This suggests that other known inductive currents should be incorporated and/or a more complex conductivity structure than the 5-layer 1-D model used here should be implemented for this region.

We have created a database consisting of hourly means of the geomagnetic field components observed on quiet days in years 2001–2002 on ground observatories, and Ørsted and CHAMP satellite measurements covering the same time intervals. In the first part of our study, we use the potential method to estimate the model of external inducing field. Following 3-D simulations are used to evaluate the effect of heterogeneous surface conductance map on Ørsted and CHAMP satellite measurements and to compare the results with observations. Improvement of up to 15% with respect to the best 1-D model was observed in surface observatory data as well as in the Ørsted and CHAMP measurements.

Wavelet spectral analysis permits quantitative monitoring of the signal evolution by decomposing a time-series into a linear superposition of predefined mathematical waveforms, each with finite duration and narrow frequency content. Thus, the frequency range of the analyzing wavelets corresponds to the spectral content of time-series components. We present a wavelet analysis of 3 years of vector magnetic data from the CHAMP satellite mission. We have detected, identified and classified not only artificial noise sources (e.g. instrument problems and preprocessing errors) but also high frequency natural signals of external fields (e.g. F-region instabilities and pulsations). The results of this analysis will be used for: (a) consequent correction and flagging of the data, (b) derivation of a clean (undisturbed) dataset suitable for the purposes of crustal and main field modeling, and, (c) study of natural signals (e.g. F-region instabilities, pulsations) contained in the data.

Recently Olsen & Kuvshinov (2003) presented an approach for modelling the ocean effect of geomagnetic storms by solving the induction equation given the conductivity distribution of the Earth's interior and the time-space structure of the storm. The results for several major storms show much better agreement between the observed and the simulated magnetic vertical component at coastal sites if the oceans are considered and contributions from spherical harmonics other than

P

{sk1/0} are included. In that paper we report on the results for selected ground observatories. Here we investigate the time-space distribution of the anomalous induction (ocean) effect of geomagnetic storms at ground as well as at satellite altitudes in order to answer the following question: Where and during which phase of the storm can we expect to observe the ocean effect from ground and from satellite altitudes, and what is the expected amplitude? Our simulations show that during the main phase of major geomagnetic storms the ocean effect in the radial component,

B

r

, and in the scalar field,

dF

, should be clearly visible at CHAMP and Ørsted altitudes and might reach magnitudes of tens of nT. The effect is especially prominent near the Pacific coast of North America, and near the southern edges of Africa, Australia and South America.

Recently Tyler et al. (2003) demonstrated that the magnetic fields generated by the lunar semidiurnal (M

2

) ocean flow can be clearly identified in magnetic satellite observations. They compared their numerical simulations of magnetic fields due to the M

2

tide with CHAMP observations and found close agreement between observations and predictions. Their three-dimensional (3-D) conductivity model consists of a surface thin shell of variable conductance and an insulating mantle underneath. Some discrepancies between observations and predictions have been addressed to the absence of a coupling between the surface shell and the mantle. Here we performed model studies of the magnetic signals due to ocean tidal flow in order to answer the following questions. (1) How does the inclusion of a conducting mantle affect the magnetic signals of the M

2

tide at CHAMP altitude? (2) Are magnetic signals from other tidal components detectable at CHAMP altitude? (3) What amplitude has the magnetic M

2

tide at Ørsted altitude? The 3-D conductivity model that we consider incorporates a thin shell and either a radially symmetric or a 3-D mantle underneath. Our model studies demonstrate that including a conducting mantle yields significant changes of the magnetic M

2

oceanic signals, with peak-to-peak values at CHAMP altitude of order 3nT. The magnetic signals due to other prominent ocean tidal modes (like K

1

and O

1

) are below 0.5 nT at CHAMP altitude. The M

2

peak-to-peak magnetic signal at Ørsted altitude is of order 1 nT.

The CHAMP satellite observed large enhancement of the neutral atmospheric density in the southern auroral region during the Sept. 25–26, 2001 magnetic storm. This enhancement was accompanied by intense auroral electrojets and strong small-scale field-aligned currents. Joule heating rate and the heating rate by these small-scale field-aligned currents are compared. Results show that the heating rate by small-scale field-aligned currents is about 2∼3 orders of magnitude higher than that by the Joule heating, indicating its importance as a heating source.

The comprehensive data from Ørsted and CHAMP adding to the data base from the earlier Magsat magnetometry mission have made it possible to develop sophisticated models for the high-latitude field-aligned currents (FAC) relating to solar wind and magnetospheric parameters. However, there are still large uncertainties involved in using the new models primarily due to the lack of precise knowledge of the temporal and spatial development of the currents in relation to the highly variable solar wind parameters and the still unpredictable effects of magnetospheric substorms. A further problem is the lack of proper account of fine-structure currents in the statistical models. By considering separately the upward and downward FAC intensities the paper presents some first attempts to examine the possible diminishing effects on modelled FAC magnitudes from averaging oppositely directed currents.

The first SIRCUS campaign covered the days February 16–22, 2002. It involved the CHAMP and Ørsted LEO satellites and the EISCAT (Tromsø, Svalbard) and Sondrestrom incoherent scatter radars. The primary data set was supplemented by DMSP-F13 and -F14 particle flux and ACE solar wind measurements. On February 21 all instruments were operated in campaign mode and delivered usable data. We identify the low-altitude cusp based on signatures in our multi-instrument data set and derive a consistent time-dependent mapping of the cusp in terms of magnetic local time and latitude. We demonstrate that small-scale variations of the magnetic field resp. field-aligned currents (spatial scales of several hundred meters) can be used to identify the cusp. However, their general ability as cusp signature has not yet been proven. An extended analysis of several SIRCUS campaigns is expected to give a qualified answer.

The Ørsted high-precision magnetic measurements made at high temporal resolution of 25 and occasionally 100 samples/sec corresponding to spatial resolutions down to less than 100 m have demonstrated the occurrences of highly structured field-aligned currents (FAC) in the low-altitude Cusp region. The observed magnetic perturbations indicate structures of very intense but thin sheets or narrow filaments of mixed up- and downward currents up to several hundreds of µA/m

2

embedded in large-scale FAC structures of only up to a few µA/m

2

. The intensities and locations of fine-scale FAC structures are closely related to solar wind conditions.

We present initial results of a comparative study of Pi2 pulsations observed by the CHAMP satellite and at the Sutherland ground station [32° 24′ S, 20° 40′ E]. Times when a Pi2 pulsation was observed on the ground (predominantly during night-time) and when CHAMP was located within 30° of longitude of Sutherland and at latitudes less than 50° N and S were selected for study. Following pre-processing and inspection to exclude unsuitable events, the satellite vector magnetic field data were rotated into a field aligned coordinate system and band-pass filtered in the Pi2 frequency band (typically .005 – .05 Hz).

We present the analysis of a Pc 3 geomagnetic pulsation event observed simultaneously by CHAMP and by the South European GeoMagnetic Array SEGMA (1.56 <

L

< 1.88) during the conjunction of July 6, 2002. Both compressional and transverse oscillations were identified in CHAMP magnetic measurements. A close correspondence between the compressional component and the ground signals is observed. At the same time the joint analysis of space and ground observations clearly indicates the occurrence of a field line resonance at

L

✠ 1.6. A direct confirmation of the well known 90° rotation of the ULF wave polarization ellipse through the ionosphere is also provided.

The equatorial electrojet signal recorded on the Overhauser Magnetometer (OVM) on board CHAMP can be divided into four classes: the classical electrojet (EEJ), the counter electrojet (CEJ), no electrojet (NEJ) and a new term, formation of the electrojet (FEJ). A statistical study on the location, period and corresponding magnetic events shows that at longitudes near 60°, 165°, 255° and 345° the morning events are CEJ, which then slowly form into FEJ near noon, and finally, as local time increases, the full EEJ is realized. All other longitude zones have only EEJ events. The reason for the occurrence of these events at certain longitudes is as yet unclear.

ESPERIA is an equatorial space mission mainly conceded with detecting any tectonic and preseismic related signals. More in general, it has been proposed for defining the near-Earth electromagnetic, plasma, and particle environment, and for studying perturbations and instabilities in the ionosphere-magnetosphere transition region. The same multi-instrument payload also allows anthropogenic electromagnetic emissions to be investigated. To study earthquake preparation processes and anthropogenic impacts in the Earth's surface, a phase A study has been realized for the Italian Space Agency. Within this framework also the ARINA particle experiment is in progress.

The routine processing of the CHAMP data has started in May 2001. Since then calibrated data, (science quality) Level-2 products, are made available to the community through our data centre ISDC. The gained experience with the behaviour of the instruments prompted some small adjustments of the processing codes. Also feedbacks from the users were considered when appropriate. Here we will report on the availability and quality of the data on an instrument-by-instrument base. Within the Magnetic and Electric field (ME) group we process the scalar and vector magnetic field data, evaluate the star camera readings and the ion drift-meter measurements. For a complete cross-calibration between the scalar and vector magnetometers two dedicated attitude manoeuvres have been performed. A modified processing of the star camera readings has reduced the attitude noise during periods of single head solutions. As a rather recent product, we make the local electron density available. It is planned to estimate also the electron temperature. Both quantities are derived from the Planar Langmuir Probe.

The 1980s and 1990s saw the Global Positioning System (GPS) transform space geodesy from an elite national enterprise to one open to the individual researcher. By adapting the tools from that endeavor we are learning to probe the atmosphere and the ocean surface in novel ways, including ground-based sensing of atmospheric moisture; space-based profiling of atmospheric refractivity by active limb sounding; and global ocean altimetry with reflected signals. Ground-based GPS moisture sensing is already being tested for weather prediction. Limb sounding is less mature but offers a variety of attractions, including high accuracy, stability, and resolution; all-weather operation, and low cost. GPS “reflectometry” is least advanced but shows promise for a number of niche applications.

We showed that the amplitude of GPS occultation signal is important indicator of the ionospheric activity. Amplitude is more sensitive to small-scale ionospheric disturbances than the phase of the radio occultation (RO) signals. Local mechanism of strong ionospheric influence on the amplitude and phase of RO GPS signals is described. Critical points (tangent points) in the ionosphere, where the gradient of the electron density is perpendicular to the RO ray trajectory, strongly influence on the amplitude and phase of RO signals, and introduce multi-RO ionospheric effect in the experimental RO data. Positions of the critical points depend on the structure of the ionospheric disturbances. Analytical model for description of multi-RO ionospheric effect is introduced. Model accounts for the horizontal gradients in the ionosphere and gives analytical expressions for the phase path excess and refraction attenuation of the radio wave propagating through the disturbed ionosphere. Analytical model and analysis of the CHAMP RO data indicated that the centers of strong ionospheric influence on RO signals can exist, for example, in the sporadic E-layers inclined by 3–6 degrees relative to the local horizontal direction. In this case one can observe simultaneously with atmospheric RO the appending ROs in the ionospheric layers. Multi- RO effect can be a cause of the ionospheric interference in the communication and RO signals. Multi- RO effect can be used to study the structure of the ionospheric disturbances using the amplitude variations in RO signals. Multi- RO effect allows introducing a classification of the ionospheric influence on RO signals using the amplitude data. This indicates a possibility for separating the regular and random parts in the ionospheric contribution in the RO signals.

The GPS radio occultation technique onboard LEO satellites such as CHAMP is a rather simple and relatively inexpensive tool for profiling the ionospheric electron density from satellite orbit heights down to the bottomside. The paper addresses capabilities of the ionospheric radio occultation (IRO) technique for globally monitoring the ionosphere on a routine basis to derive value added data products and study various ionospheric processes including perturbations. A model assisted retrieval technique, operational data processing and validation of vertical electron density profiles are also discussed. These profiles may not only be used to validate ionospheric models, they provide also a good data basis for developing new models of key ionospheric parameters such as the critical frequency foF2, the peak height hmF2 and the scale height Hs. Such models are helpful to improve retrieval procedures and tomographic reconstruction techniques. Due to the operational data processing capabilities the data products may contribute to space weather monitoring.

Ionospheric radio occultation (IRO) measurements have a big potential for monitoring the ionospheric behavior on global scale for now- and forecasting the ionospheric impact on radio systems. In this article we validate the retrieved vertical electron density profiles (EDPs) derived from IRO measurements onboard CHAMP by using vertical sounding measurements at five European vertical sounding stations — Athens, Dourbes, Juliusruh, Rome and Tortosa. Since first IRO measurement onboard CHAMP in April 2001, more than 70000 electron density profiles have been retrieved by a model assisted technique so far. The comparison of IRO retrieved EDPs with ionosonde profiles obtained from the above mentioned stations will be discussed.

Abel inversion offers a straightforward way to obtain the vertical distribution of electron density with low computational load. Nevertheless the treatment of the electron density above the LEO orbit must not be neglected, specially for satellites with very low orbit such as CHAMP. This work extends previous results obtained by inverting real GPS data from LEO data, coming from satellites such as CHAMP or SAC-C. In this work, the topside ionosphere is modelled using positive elevation data. To overcome the spherical symmetry assumption, occultations are processed with the aid of Vertical Total Electron Content, estimated from ground GPS data or models. The resulting electron density profiles are compared with external real data consisting basically on basic parameters or true-height vertical profiles obtained from ionosonde measurements.

Presented are first results in retrieving, analysing, and modelling the topside plasma scale height by using CHAMP ionospheric radio occultation observations. The plasma scale height value in the region situated immediately above the ionospheric F2-layer density peak, is very important for the TEC calculation and plasma density reconstruction procedures based on GPS radio occultation measurements. Based on the year-long time series data accumulated so far, obtained were latitudinal, diumal, and seasonal variations of the topside plasma scale height. Considering the growing CHAMP measurement data base, it is possible to develop a new empirical model to be used for improving the process of electron density profile retrieval by delivering an improved initial guess of the topside electron density profile.

The knowledge of the transmitter and receiver differential code biases (DCB) plays a key role for the calibration of GPS based measurements of total electron content (TEC). To estimate the DCB of the CHAMP receiver concerning the zenith looking antenna a model assisted technique has been developed which takes advantage of the known GPS biases and comparatively low TEC above Low Earth Orbiter (LEO) altitudes in polar regions and during nighttime intervals. For assistance the Parameterized Ionospheric Model (PIM) is used. Applying this method we derived a reliable bias solution for the CHAMP receiver varying within a spread RMS below 1 TECU for the years 2001 and 2002. The SAC-C receiver DCB (zenith antenna) could be derived comparatively stable. In this paper we describe the model assisted calibration technique and present differential code bias estimations for CHAMP and SAC-C. The model assisted approach has been used to estimate the CHAMP occultation DCB as well. This bias could also be assessed by simultaneous GPS observations of the zenith and occultation antenna for selected intervals in 2001. DCB estimations from both approaches are shown.

Both the CHAMP and SAC-C satellites permanently track several GPS satellites using dedicated zenith looking antennas for precise orbit determination. These 0.1 Hz sampled dual frequency navigation measurements provide valuable information on the ionization state of the upper ionosphere and plasmasphere up to GPS altitudes on a global scale. After preprocessing and calibration, link related TEC measurements are derived from the GPS navigation observations. Three-dimensional electron density information is retrieved by assimilating these TEC data into the Parameterized Ionospheric Model (PIM). In specific periods CHAMP and SAC-C circle the earth in nearly the same orbit plane due to their different orbit properties. We focus on these constellations which are of special interest for the combined assimilation of TEC measurements from both satellites. Reconstruction results for selected assimilation examples are presented by means of two dimensional slices along the respective CHAMP/SAC-C orbit plane. Using electron density measurements from the CHAMP Langmuir Probe we validate our reconstruction results along the CHAMP path and discuss the impact of the TEC measurements from SAC-C on the reconstruction result at CHAMP orbit altitudes.

Integrated electron density measurements along GPS signal ray paths provide a useful data source for 3-dimensional ionospheric reconstruction. Since a substantial data amount is required for the application of the inverse problem, ground-based inversions are restricted to regions with high ground receiver density. The additional incorporation of space-based GPS data from LEO satellites allows the monitoring in areas less densely covered by ground receivers. We present a case study of 3-dimensional reconstruction for Nov. 2001 in the North polar region. Input data are calibrated TEC derived from IGS and CHAMP measurements. The results show increased electron densities near the geomagnetic pole. The comparison of the reconstruction results with measurements of the Langmuir Probe onboard CHAMP shows a good agreement along the CHAMP orbit. The large-scale structure of the ionospheric observations are closely related to the present geomagnetic conditions.

Vertical electron density profiles (EDP) derived from ionospheric radio occultation (IRO) measurements onboard the German CHAMP satellite are calculated by the Institute of Communications and Navigation of DLR on a regular basis since 11 April 2001. In order to validate the ionospheric radio occultation data obtained from this satellite, any systematic comparison with independent measurements but also with results from ionospheric models should help to get a better feeling about the quality of the data. On the other hand, if the IRO data quality is sufficient, the data may contribute to evaluate the accuracy of the ionospheric models. In this study we compare CHAMP EDP data derived in 2002/03 with the NeQuick model, which is one of the three electron density models developed at the Abdus Salam ICTP in Trieste (Italy) and the Institute for Meteorology and Geophysics in Graz (Austria). We discuss results of this comparison, showing changes in bias and deviation with latitude and local time. The best agreement between both types of data was found above a height of 300 km.

Rising accuracy requirements for attitude and translation control systems of satellites require better models for the expected disturbance forces. A major non-conservative disturbance force in Low Earth Orbits (LEO) is the atmospheric drag. With current atmospheric models an estimation of the drag force can be calculated, however these atmospheric density models, like the NRLMSISE-00[3] model, do not include short-term density variations. These density variations can no longer be disregarded. This paper deals with the development of a model that does include these short-term density variations.

Atmospheric sounding with the German CHAMP (CHAllenging Minisatellite Payload) satellite is successfully performed since February 2001. In total ∼145,000 precise globally distributed vertical profiles of refractivity, temperature and water vapor were provided as of April 2004. The operational occultation infrastructure from GFZ allows for the demonstration of Near-Real Time (NRT) data analysis since February 2003. An average delay of ∼5 hours between each measurement and provision of corresponding analysis results is continuously reached. A comparison with more than 10,000 radiosonde measurements shows nearly biasfree refractivity and temperature between ∼7 and ∼30 km. The standard deviation is ∼1 % and ∼2 K, respectively. Data of the SAC-C (Satelite de Aplicaciones Cientificas-C) occultation mission are used to prepare for multi-satellite capability of the operational data analysis system. Future prospects of the provision of occultation analysis results at GFZ are given.

We present an Optimal Estimation processing of simulated radio occultation data. Temperature and water vapor profiles are retrieved from either bending angle or refractivity measurements. The advantages of one over the other are assessed. A comparison with idealized calculations shows the potential of ray tracer calculations.

The canonical transform (CT) and full spectrum inversion (FSI) method together with their heuristic sliding spectral modifications are validated using end-to-end simulation data and one week of CHAMP observations. In general, we observe a pronounced correlation between small refractivity biases and enhanced penetration altitudes. Processing of simulated occultation data shows that the heuristic methods exhibit smaller retrieval errors vindicating the assertion that the sliding spectral approaches react less sensitive to receiver tracking errors. The corresponding mean retrievals errors found in the CHAMP data analysis, however, are consistent within 0.5%; differences are observed with respect to penetration altitudes and the retrieval errors' standard deviations.

Diffractive integrals use a reference signal to derive the field distribution from the radio holograms. The phase of reference signal coincides with phase of the Green function and can be found by solving the scalar wave equation. This property of the reference signal demonstrates a fundamental nature of the radio holographic focused synthetic aperture principle (RFSA) specifically in its application to the radio occultation (RO) data analysis. The RFSA method has been validated by direct observations of multi-beam propagation in the atmosphere and weak reflections from the terrestrial surface, its vertical resolution has been estimated early as about 70 m. To compare the Canonical Transform (CT), Back Propagation (BP) and RFSA methods a general inverse operator (GIO) is introduced. CT as a partial case of GIO method can resolve physical rays in multi-path situations under an assumption of the global spherical symmetry of the atmosphere and ionosphere. RFSA method can account for the multi-path in the case when the global spherical symmetry is absent by using the appropriate model of the refractivity and has a promise to be effective for operational data analysis.

We discuss the application of canonical transform (CT) methods for analyzing radio occultations with multipath behavior. In the present work we discuss a mapping to the representation of approximate impact parameter. This method generalizes the full spectrum inversion (FSI) method for the case of noncircular orbits. The method is based on mapping the field by a Fourier integral operator that maps directly from the measured time-dependent field to the impact parameter representation without first doing a back propagation. Furthermore, our method allows for simple, asymptotic direct modelling of wave propagation. We perform processing of simulated radio occultations and show that our method provides the same accuracy and resolution as the standard CT method.

A major application of the CHAMP occultation data is the preparation of processing systems for future occultation missions. The University of Graz (IGAM) uses data and analysis results to prepare for the ACE

+

multi-satellite occultation mission, which is currently foreseen to be launched in 2008. We compare vertical profiles of refractivity, derived by GFZ and IGAM, with ECMWF data (European Centre for Medium-Range Weather Forecasts) and discuss the deviations. Good agreement is observed in the upper troposphere and stratosphere. The magnitude of the observed refractivity bias in the lower troposphere depends significantly on the quality control criterion. The bias can be nearly eliminated by the application of the Full Spectrum Inversion analysis method down to ∼1 km above the Earth's surface.

Two radio occultation (RO) retrieval schemes designed at IGAM to enhance the performance at high altitudes are presented, applied to CHAMP RO data, and validated against ECMWF analyses, GFZ operational retrieval, ENVSAT/MIPAS-, and ENVIAT/GOMOS-derived temperature profiles. IGAM proposes to include background information into the RO retrieval only at one point at bending angle level in order to be able to track error characteristics of the retrieved product. The results show very good agreement with GFZ retrieval and ECMWF analysis below 15 km and, depending on the background information used, either a significant warm bias or essentially no bias up to 30 km. Compared to MIPAS, the only independent data source, the IGAM/ECMWF retrieval is unbiased up to 40 km.

The present work investigates the possibility of retrieving humidity profiles using the bending angle data obtained from GNSS Radio Occultations (Global Navigation Satellite System RO) without data constraint. In particular with the proposed approach, the dry pressure profiles are obtained by fitting the bending angles of the outer troposphere layers (from h=h

250K

up to the stratopause) using the Hopfield dry atmosphere model. In this model the ground pressure and temperature are the parameters to be estimated. The humidity profiles are extracted by subtracting the contribution due to the dry atmosphere from the measured bending angles. In the first part we discuss the mathematical approach adopted for the derivation of refractivity profiles without using the Abel inversion and water vapor directly from the bending angle. In the second part the results are shown from applying the method on CHAMP data. In particular, we have used refractivity profiles from CHAMP which were derived using a heuristic retrieval algorithm based on the canonical transform.

The developments of an operational system for processing of radio occultations from the GRAS (Global navigation satellite system Receiver for Atmospheric Sounding) receiver onboard the future Metop satellite is the primary scope of the EUMETSAT project: GRAS Satellite Application Facility (SAF). The planned GRAS SAF data products and processing system is discussed together with a preliminary product validation based on statistical analysis of retrieved refractivity and temperature using radio occultation data from the German CHAMP (CHAllenging Minisatellite Payload) satellite. The CHAMP data from the first 8 weeks of 2003 have been processed using the GRAS SAF retrieval system and we present the results as compared to the ECMWF. The mean difference in refractivity is approx. 0.5% with standard deviations less than 1% between 8–25 km and approximately 2% both close to the surface and at 35 km.

The analysis of the amplitude of the GPS/MET and CHAMP radio occultation (RO) events revealed clusters (quasi-regular structures) with a vertical size of about 10 km and an interior vertical period of ∼0.8 to 2 km in the tropopause and lower stratosphere. The height interval of the clusters changes from 10 to 40 km. Restored from the RO amplitude data, variations of the vertical temperature gradient in clusters dT/dh are from −8 to −9 K/km up to 6 to 8 K/km. These variations can be associated with the influence of internal waves (gravity waves; GWs) propagating through the atmosphere and mesosphere. We found a height dependence of the GW phase and amplitude (the GW “portrait”), using for example the amplitude data corresponding to the GPS/MET and CHAMP RO events. For a GPS/MET event we estimated the horizontal wind speed perturbations which are in fairly good agreement with radiosonde data. Between 10 and 40 km, the horizontal wind speed perturbations v(h) are changing in the range v from ∼±1 to ±9 m/s with vertical gradients dv/dh ∼±0.5 to ±15 m/(s km). The height dependence of the GW vertical wavelength was derived by differentiation of the GW phase. The analysis of this dependence led to the estimation of the GW intrinsic phase speed which changed for the considered events in the interval from 1.5 to 5 m/s. The analysis of RO data shows that the RO signal amplitude contains valuable information for studying the GW activity in the atmosphere.

Global analyses of E

p

as a proxy for gravity wave activity in the stratosphere obtained using GPS radio occultation (RO) data from the CHAMP satellite are presented. Large E

p

values are noticed at tropical latitudes, but also at midlatitudes during winter. A correspondence between tropical gravity wave activity and tropical deep convection is found. E

p

values are also large during the sudden stratospheric warming that occurred in the southern hemisphere during September–October 2002.

In this paper an overview of the temperature structure in the tropical upper troposphere and lower stratosphere (UTLS) region is given using Global Positioning System (GPS) radio occultation (RO) data from the German CHAMP (CHAllenging Minisatellite Payload) satellite mission. Several climatologies for tropopause parameters based on radiosonde data and model analyzes have been published in recent years. Both the data sources suffer either to less global coverage or low vertical resolution. This fault can be overcome by the GPS RO technique due to its global coverage, high vertical resolution, all-weather viewing, and long-term stability. CHAMP RO data are available since 2001 with up to 200 high resolution temperature profiles per day. Using CHAMP RO data during May 2001-September 2003, the structure and variability of the tropical tropopause is presented.

The temperatures retrieved from MIPAS/ENVISAT limb mid-infrared emission and CHAMP GPS radio occultation measurements are compared at altitudes between 8 – 30 km during the stratospheric major sudden warming in the southern hemisphere winter of 2002. The mean differences between the correlative measurements of the two instruments are less than ∼1 K with rms deviations of ∼3–5 K. The MIPAS temperatures are slightly higher than those of GPS-RO around 30 km. Possible explanation is discussed.

This analysis presents comparisons of the atmospheric temperatures retrieved from GPS/SAC-C radio occultation observations using the JPL retrieval software, and from MIPAS/ENVISAT infrared spectrum measurements using the IMK data processor. Both individual profiles and zonal means of the atmospheric temperature at different seasons and geo-locations show reasonable agreement. For the temperatures at altitudes between 8–30 km, the mean differences between the correlative measurements are estimated at less than 2 K with rms deviations less than 5 K. A similar cross comparison technique can be used to help validate the observed temperatures from the new EOS MLS instrument, to be launched in 2004.

The global structure and variability of the tropopause observed using CHAMP/GPS radio occultation observations from May 2001 through September 2003 are presented. Tropopause temperature and height observed by CHAMP/GPS has been compared with radiosonde observations at a sub-tropical site. At polar latitudes, the tropopause sharpness found to be highest in summer and lowest in winter with slight differences between hemispheres.

A data assimilation process that employs best linear unbiased estimation has been applied to slant total electron content data from the IGS network and CHAMP. The data has been assimilated into a three dimensional electron density grid for a test day in June, 2003. The accuracy of these assimilations has been assessed through the use of independent GPS data.

GPS radio occultation phase delay (PD) data contain valuable information about atmosphere and ionosphere and its structure, based on the recorded changes of the direct GPS signal between LEO and GPS satellite. PD data contain signatures of surface reflections mainly above water and snow/ice covered regions which have successfully been detected and analyzed with the radio holographic (RH) method recently. Reflection signatures in the PD data can also be detected and analyzed with the continuous wavelet transform (CWT) in a strait forward manner, e.g. without requiring any additional information/forward model or reference field. The use of the CWT as an additional tool to analyze PD data is described and the assets and drawbacks in comparison to RH are discussed. Two years of consistent CHAMP PD data have been analyzed with CWT. The signature of surface reflections can be detected successfully. The CWT based method also reveals weak signatures during times when the direct GPS signal is not influenced by the atmosphere and travels only through the ionosphere. Different classes of ionospheric signatures can be isolated. The geographical distribution of each signature class reveals different patterns. Such observations may contribute to investigations of ionospheric structures like sporadic E or spread F layer.

In this paper a description of the CHAMP atmospheric processing system for radio occultation data at GFZ Potsdam is given. The generation of radio occultation products, as e.g. atmospheric excess phases, vertical profiles of refractivity, temperature or water vapour is a complex process. Besides the scientific challenge the design and installation of an automatic data processing system is also of great importance. This system must be able to process the different input data from external data sources, coordinates the different data streams and scientific software modules, and feeds the results into the data centre automatically. Caused by different user demands the CHAMP Atmospheric Processor is divided into two parts: A rapid processing mode makes radio occultation analysis results available on average five hours after measurements. In the standard processing mode quality checked profiles of atmospheric parameters are available with a latency of about two days.

This paper studies one aspect of use of the CHAMP GPS occultation data, namely the improvement of the atmospheric pressure field, particularly over Antarctica. Previous studies indicate that pressure differences between ECMWF and ground truth data reach 5.18 hPa RMS over Antarctica [Ge et al., 2003]. In this study, comparisons of pressure profiles (January–March 2003) from data (CHAMP occultation and radiosonde) and models (ECMWF and NCEP), indicate large discrepancies over different regions, notably over southern polar region. Global pressure differences between CHAMP and radiosonde and model outputs reach 4 hPa RMS at 1 km above MSL. We found a positive bias in CHAMP data (CHAMP measures larger pressure values) when comparing with both radiosonde and ECMWF. Analysis shows the lack of adequate penetration of CHAMP occultation data in the planetary boundary layer particularly in the tropical region (only ∼10% signal is within 1 km above MSL), as compared to ∼80% penetration in Arctic and Antarctica. However, CHAMP provides improved data coverage in temporal, spatial and vertical resolution globally. We conclude that the CHAMP occultation data could potentially improve the surface pressure modeling to benefit temporal gravity recovery, in particular over data sparse region such as Antarctica.

The gravity wave activity on a global scale between 10 and 27 km was analyzed by using temperature profiles retrieved from GPS occultation events detected with SAC-C and CHAMP satellites. Examples of its variability with altitude, longitude, latitude and time are presented for different wavelength ranges. A particular attention is given to gravity wave generation in the Andes Range region, in connection with the forcing of the mean wind by the mountains, and in the equatorial region, in possible connection with convective activity. A clear signature, possibly related to mountain wave generation, is seen above Andes mountains, between 30 and 40S. It is also appreciated the global enhanced wave activity in equatorial regions during the winter hemisphere.

The CHAMP radio occultation (RO) data provide the first opportunity to create real RO based climatologies on a longer term. CHAMPCLIM is a joint project of the Institute for Geophysics, Astrophysics, and Meteorology (IGAM) in Graz and the GeoForschungsZentrum (GFZ) in Potsdam. The overall aim of CHAMPCLIM is to ensure that the CHAMP RO data are exploited in the best possible manner, in particular for climate monitoring. The main objectives of the CHAMPCLIM project can be summarized in form of three areas of study as follows: RO data processing advancements for optimizing climate utility, RO data and algorithms validation based on CHAMP/GPS data, and global RO based climatologies for monitoring climate change. Here we show a summary of the current activities and exemplary results.