G31A-0780 0800h
Combination of VLBI, GPS, SLR, DORIS and LLR Data at the Measurement Level: EOP and TRF computations
In the framework of activities of the International Earth Rotation and Reference Systems Service (IERS) Combination Research Centres (CRC), the french {\it Groupe de Recherche en G\'eod\'esie Spatiale (GRGS)} studies the benefit of combining several geodetic techniques (SLR, LLR, VLBI, GPS and DORIS) at the measurements level in order to obtain a global and consistent solution for Earth Orientation Parameters (EOPs): polar motion $xp$ and $yp$, universal time $UT1$ and celestial pole offsets in longitude and obliquity $d\psi$ and $d\epsilon$ with a six-hour sampling, as well as weekly station positions. A one-year test period (the year 2002) has been chosen to prove the power of such a combination moreover worked out in a homogeneous global terrestrial reference frame. All techniques were processed with the same computational framework so with the same \textit{a priori} models and \textit{a priori} values for parameters. The optimal relative weights between each geodetic technique were obtained with the Helmert's optimal variance component estimation. The aim of this paper is to describe the processing and the precision/resolution level of each individual technique for EOPs and stations positions, to show how we handled with the combination of techniques, and to discuss results for EOPs by comparing them to the individual solutions per technique or to other external solutions.
G31A-0781 0800h
Combination of GPS and DORIS Data for Troposphere and Local tie Studies
Combining satellite geodetic data at the measurement level is a big challenge as several types of systematic errors can arise and could forbid any common processing scheme or at least bias the derived geodetic results. However, the global assimilation of these different data could help us increase the accuracy of common parameters or allow us to derive more frequent determination of EOP parameters. In a first attempt of a real multi-technique combination, we have processed GPS and DORIS ground data during the CONT-02 campaign (October 2002). GPS orbits and clocks were held fixed while all DORIS satellites parameters and stations coordinates were estimated in a free-network adjustment. In second step, we have added GPS data from the TOPEX/Poseidon mission, making full-use of one these rare collocations in space. At ground collocated sites a unique tropospheric parameter was estimated as a randomwalk process plus an unknown bias. Stations coordinates were estimated as well an unknown local tie vector. All DORIS orbits (including TOPEX) were estimated simultaneously. The impact of such GPS+DORIS combination was investigated at collocated sites. In particular, derived tropospheric solutions were compared internally during daily overlap periods and externally with VLBI results when available. Observed biases and potential accuracy improvements are discussed. Local ties vectors were also estimated and compared to current ITRF determinations to investigate the use of such technique to derive "pseudo-local ties" at all collocated sites.
G31A-0782 0800h
Tropospheric Parameters and Subdaily EOP From Combinations of Independent Space Geodetic Data
The space geodetic techniques GPS, VLBI, SLR and DORIS contribute to the determination of several geodetic parameters (e.g. site positions, Earth orientation parameters (EOP), tropospheric parameters) providing valuable information to study various geophysical processes. Due to the different strengths of the techniques it can be expected that the parameters benefit from a combination. The VLBI campaign CONT02, initiated by the IVS, provides 15~days of continuous VLBI measurements. Therefore, this data set is well-suited for the combination with other techniques. Especially the combination with other microwave techniques like GPS provides the opportunity to estimate common tropospheric parameters in addition to station coordinates and EOP. For the studies presented here, free daily normal equations were generated for GPS and VLBI using identical models and the same parameterization to avoid any inconsistencies. Additionally, the normal equation of a 14-day SLR solution is included to investigate primarily reference frame related aspects. The work focusses on the combination of tropospheric parameters and EOP with a high resolution in time: solutions with one and two hour resolution of the parameters were compared to decide whether a higher time resolution is more appropriate to describe the time-dependent behavior of these parameters. For the validation of the tropospheric parameters independent data sets of water vapor radiometers are used, and the EOP are compared with a subdaily model derived from altimetry. Special attention has to be addressed to the tropospheric parameters from GPS, because they are sensitive to the physical characteristics of the antenna and the antenna environment. The comparison with VLBI-derived tropospheric parameters shows that absolute antenna phase center corrections should be used instead of relative models. Similarly, if a radome is installed at the antenna, the tropospheric zenith delay estimates change significantly. As no phase center calibrations are available for antenna/radome configurations, this effect limitates the benefit to be achieved by a combination. However, the results show, that a rigorous combination improves the quality of all estimated parameters compared to the single technique solutions. In addition to the microwave techniques, SLR normal equations were included in the combination studies as well, but only with a daily resolution for the ERP. The advantage of SLR must be seen in the contribution to a more stable definition of the reference system due to more co-located stations and a better realization of the geocenter. Furthermore, SLR can deliver valuable information for the scale of the combined network.
G31A-0783 0800h
Power Spectral Analysis of Simultaneous VLBI and GPS Tropospheric Estimates
Observations by space geodetic techniques experience refraction and signal delay due to passage through the Earth's atmosphere. For high-accuracy positioning results, data analysts must account for these effects. Since independent path delay values of sufficient accuracy are not usually available, nuisance parameters are commonly added in the geodetic analysis. The general validity of such zenith path delay (ZPD) estimates as true atmospheric measures has been confirmed by comparison of results from independent radiometric and other techniques over many years. Biases and standard deviations in the sub-cm range are normally found, which is expected to be adequate as inputs to improve the forecast performance of numerical weather models. To better understand the noise characteristics of ZPD estimates from VLBI and GPS, we have examined the power spectra of simultaneous observations during a 15-day period in October 2002. The official combined ZPD products from the technique services have been used primarily, but series from individual analysis centers have also been included. For the seven sites studied, the power-law spectral indices over sub-daily intervals are close to -8/3, consistent with fully developed Kolmogorov turbulence, and flatten over longer periods. The VLBI series, sampled hourly, show white noise at levels of 0.7 to 1.5 mm for frequencies above 5 cycles per day. The simultaneous GPS series, sampled every 2 hours, display no indication of white noise except for one receiver with poor data analysis. The spectra of VLBI-GPS differences are generally flat but show possible signs of excess noise in some spectral bands. Based on these results, estimating VLBI ZPD values more often than every few hours should be reconsidered, especially if changes would strengthen other parameters. On the other hand, GPS-based ZPD estimates should be determined more frequently, at least hourly. Considering the greater reliability of the VLBI scale and the corresponding weakness of GPS (mostly due to uncertainties in the phase patterns of the satellite transmit and tracking station antennas), an optimal combination of ZPD results should rely on the VLBI estimates for its absolute long-term accuracy and weight GPS heavily for its high-frequency precision.
G31A-0784 0800h
Validation of Atmospheric Refraction Modeling Improvements in Satellite Laser Ranging.
Atmospheric refraction is an important accuracy-limiting factor in the use of satellite laser ranging (SLR) for high-accuracy science applications. In most of these applications, and particularly for the establishment and monitoring of the TRF, of great interest is the stability of its scale and its implied height system. The modeling of atmospheric refraction in the analysis of SLR data comprises the determination of the delay in the zenith direction and subsequent projection to a given elevation angle, using a mapping function. Standard data analyses practices use the 1973 Marini-Murray model for both zenith delay determination and mapping. This model was tailored for a particular wavelength and is not suitable for all the wavelengths used in modern SLR systems. Mendes et al., [2002] pointed out some limitations in that model, namely as regards the modeling of the elevation dependency of the zenith atmospheric delay (the mapping function component of the model). The mapping functions developed by Mendes et al. [2002] represent a significant improvement over the built-in mapping function of the Marini-Murray model and other known mapping functions. Of particular note is the ability of the new mapping functions to be used in combination with any zenith delay model, used to predict the atmospheric zenith delay. Mendes and Pavlis [2002] concluded also that current zenith delay models have errors at the millimeter level, which increase significantly at 0.355 micrometers, reflecting inadequacy in the dispersion formulae incorporated in these models. In a next step therefore, a more accurate zenith delay model was developed, applicable to the range of wavelengths used in modern SLR instrumentation (0.355 to 1.064 micrometers), [Mendes and Pavlis, 2004]. Using ray tracing through a large database of radiosonde and globally distributed satellite data, as well as the analysis of several years of SLR tracking data, we assess the new zenith delay models and mapping functions currently available; we discuss the effect of using different types of input data to drive those models and the sensitivity of models and functions to changes in the wavelength, and we give some recommendations towards a unification of practices and procedures in SLR data analysis.
http://ilrs.gsfc.nasa.gov/science_analysis/
G31A-0785 0800h
Mitigation of Atmospheric Errors in Differential InSAR Data Using a High-Resolution Weather Model, Mauna Loa and Kilauea volcanoes, Hawaii
We investigate the utility of a mesoscale weather model for mitigating atmospheric errors in InSAR-derived displacement fields associated with volcanotectonic phenomena at Mauna Loa and Kilauea volcanoes on the Island of Hawaii. The MM5 (NCAR-Penn State Mesoscale Model Version 5) forecast is run twice daily for the island of Hawaii by the Mauna Kea Weather Center in support of Mauna Kea's astronomical observatories. MM5 has a 60-hour forecast window and the forecast fields are output in 3-hour increments. A high-resolution analysis that incorporates weather observations from National Weather Service and satellite-derived winds from the University of Wisconsin, provides the initial conditions for MM5. In turn, MM5 produces predictions of thermodynamic properties of the atmosphere, including temperature, pressure, and moisture fields at a 3 km horizontal resolution. The vertical resolution is density weighted with the greatest vertical resolution (10s of meters) near the surface. For any radar image acquisition there is a high-resolution 3D simulation of atmospheric water vapor valid within 1.5 hours of the acquisition time and predicted forward no more than 12 hours from the initial observations. Using MM5 forecast water vapor, we create line-of-sight delay maps that can be either directly removed from InSAR differential pairs or used to create synthetic radar interferograms that can be compared with the observed interferogram. We analyze Envisat ASAR radar data collected during 2003-2004 and find, qualitatively, that contours of excess path delay in MM5 model runs often closely mimic both topographic contours and fringes observed in differential interferograms which have had topographic phase removed. Over periods of days, the delays can vary in magnitude and spatial scale by 10s of mms and 10s of kms, respectively. Individual models can predict peak delays associated with moist layer inversions of order ~20 mm around the summits of Mauna Loa and Kilauea, both active volcanoes. Similar delays are also associated with the inactive Mauna Kea summit. These results are particularly pertinent for using InSAR to track accurately the history of summit-related magmatic phenomena between scene acquisitions. The path delays would introduce a significant atmosphere-related bias if included in a deformation analysis. To validate the technique, we compare total delays and precipitable water data from more than 20 continuous GPS (CGPS) sites with those estimated by the MM5 model. We investigate the technique's efficacy by inverting InSAR-derived line-of-sight displacement fields for the time-varying characteristics of a shallow magma chamber below Mauna Loa's summit caldera. Finally, we compare these results with independent inversion of summit CGPS data.
G31A-0786 0800h
On the Applicability of Taylor's ``Frozen-Flow'' Hypothesis to Spatial and Temporal Observations of Atmosphere Path Delay From InSAR and GPS
Turbulent mixing of water vapor in the lower troposphere produces fluctuations of the spatio-temporal distribution of neutral atmosphere refractive index at microwave frequencies. These variations cause phase shifts in Interferometric Synthetic Aperture Radar (InSAR) images and Global Positioning System (GPS) signals. Here, we compare spatial observations of atmospheric phase shifts from a radar interferogram of Southern California with temporal measurements of atmospheric delay obtained from a network of continuous GPS receivers operating in the imaged area. We translate temporal observations to equivalent spatial samples of delay through Taylor's ``frozen-flow'' hypothesis. We use the ``frozen-flow'' hypothesis in conjuction with Kolmogorov turbulence theory to derive theoretical expressions for temporal and spatial power spectra and structure functions of atmosphere delay. We show that temporal and spatial power spectra and structure functions have similar theoretical forms and parameters. Further, the theoretical expressions for temporal power spectra and structure functions require knowledge of the magnitude and direction of wind about each GPS receiver, which we estimate from the timeseries of delays at each site. These wind estimates and the theoretical expressions are fit to computed power spectra and structure functions of delay derived from the GPS and InSAR data. The parameters derived from the least-squares fitting of temporal power spectra and structure functions from GPS are used to infer spatial models of power spectra and structure functions of interferometric phase. Comparison of these inferred models with computed spatial power spectra and structure functions from the InSAR phase residuals demonstrate the validity of applying Taylor's hypothesis to GPS and InSAR atmospheric delay measurements. This correspondence between measurements of atmosphere delay from the two datasets suggests that full timeseries of atmosphere path delay from GPS, as opposed to delay observations taken at the radar acquisition times only, can be used to mitigate atmospheric effects in radar interferograms.
G31A-0787 0800h
Comparing And Combining Space and Terrestrial Geodetic Techniques To Monitor Crustal Deformation
It is of particular importance to define and establish appropriate observational strategies characterized by the ability to measure, with high accuracy, spatially and time continuous deformations of the land surface. An ensemble of space, airborne and terrestrial techniques can be adopted by taking advantage of the complementary strengths of the different observations. We present a network encompassing northeastern Italy, which extends from the Apennines to the southeastern Po Plain and, along the Adriatic coast, to the Venice area and Trieste. First results mainly concern a study of land subsidence performed by comparing and combining different space and terrestrial techniques, which were co-located in several stations of the network. High-accuracy knowledge of subsidence rates is of particular importance in the Venice area for the re-organization of the lagoon territory and for building an adequate protection from increasing sea level and extreme events. Examples are provided for the stations in the network such as Medicina, Bologna and Marina di Ravenna where the available data series are several years long. At Medicina, the results of CGPS, superconducting and absolute gravimeter and InSAR observations, in combination with several meteoclimatic data series, are compared and combined. At the Bologna station, CGPS, absolute gravity and InSAR data are acquired while at Marina di Ravenna, on the Adriatic coast, CGPS and InSAR observations are available. All InSAR data are processed by means of the PS technique. The ensemble of adopted space and terrestrial techniques will make it possible to acquire a thorough understanding of seasonal, long-term regional/local geophysical phenomena as well as climate-related impacts affecting the station motions.
G31A-0788 0800h
Impact of Local Gravity Observations on a Littoral Geoid
Aerogravity data collected over the Bay of Maine region were merged with existing terrestrial, shipborne, and altimetric gravity anomalies to improve the spatial and spectral characteristics of the gravity field. Previous gravimetric geoid models developed for the Conterminous United States have lacked significant data from the shoreline out to about 100 km offshore. Beyond this region, the altimetric gravity anomalies are more reliable because ocean and tide models are more accurate. Collection of data for this near-shore region in a systematic manner refined the determination of the geoid. It, thereby, enhanced discussion of the absolute accuracy of datums such as the North American Vertical Datum of 1988 as well as Sea Surface Topography. Comparison of the resulting gravimetric geoid model to available GPS/leveling data provides an independent metric for local improvements.
http://www.ngs.noaa.gov/geoid/
G31A-0789 0800h
Trends Of Vertical Crustal Displacement At Wuhan Station
GPS observation in Wuhan IGS station has been carried since 1998. The linear trend of non-atmospheric vertical crustal displacement there is determined as -6.43+/-0.20 mm/year. During the same period, gravity observation using a superconduting gravimeter has been operated in Wuhan Jioufeng station, 15 km far away from the IGS GPS station.The non-atmospheric trend predicted from the SG observation is -5.94+/-0.11 mm/year, which is almost the same as that determined by GPS.
G31A-0790 0800h
ENVISAT - Range Calibration Using a Transponder
Satellite altimetry provides a precise measure of the vertical distance of the satellite borne altimeter to the instantaneous sea surface. The accuracy of this distance depends on the calibration of the altimeter, the quality of the reflecting target and the proper estimate of path delay. Part of the altimeter calibration can be undertaken by comparison against in situ tide gauge data and GPS buoys, by inter-comparison between two altimeter data sets from concurrent satellites, or at crossing points. A different and convenient technique is the use of a dedicated transponder, a device which receives the signal from the satellite, amplifies it and re-sends it back. Opposite to the ocean surface a transponder disposes of a stable and very precise reflection reference (few millimeters), which allows for a very precise determination of the vertical distance between the satellite and the transponder. The accuracy of a such determined range depends on the ability to estimate the path delays caused by the atmosphere, the precision of the orbit and the GPS positioning of the transponder. If the dedicated transponder is operating within the footprint of the altimeter, the corresponding waveforms differ both in power and shape from the other waveforms caused by natural targets. By means of transponder signature analyses, the time of the closest approach of the satellite to the transponder is determined and the range is hence calculated. The transponder is deployed beneath the ascending Envisat track on the Greek island Gavdos. Numerical results are to be presented if relevant data are available.
G31A-0791 0800h
How VLBI Can Contribute To Ionospheric Research
Like other space geodetic techniques VLBI observations are carried out at two distinct frequencies in order to determine ionospheric delay corrections. Each ionospheric delay corresponds to the total electron content (TEC) along the ray path through the ionosphere. Because VLBI is a differential technique the observed ionospheric delays represent the differences of the behaviour of the propagation media above each two stations. Additionally an instrumental offset, independent of azimuth and elevation in which the antennas point, is contained in the measurements. Due to the sparse geographical distribution of VLBI stations and the fact that this technique is not observing every day it is suggested to derive station-specific TEC values only. These results can be used to validate the TEC values derived by other space geodetic techniques or they can be incorporated into long-term studies of the ionosphere because already now VLBI observations cover more than two complete solar cycles. A combination with GPS should improve the models, because both techniques are almost always co-located. Finally a combined long-term model of the ionosphere that uses VLBI, GPS, and satellite altimetry data back to the date of implementation of each technique will be derived. This model should be able to predict the ionosphere on longer timescales, i.e. at least over months.
G31A-0792 0800h
Combining geodetic techniques to understand the seasonal effects affecting vertical positioning
Seasonal variations are commonly observed in site position time series derived from space geodetic techniques with larger amplitude for the vertical component. These seasonal signals are mainly due to crustal deformations induced by the mass loading effects (atmosphere, ocean, snow, and soil moisture). Time-varying absolute gravity is not only due to the crustal deformation effects but is also due to the Newtonian effect (mass attraction). To better understand the observed seasonal signals and to avoid artefacts linked to the technique or to the analysis strategy, it is essential to perform a multi-technique analysis. The OCA Grasse observatory, France, operates collocated SLR and GPS space techniques, offering long time series. Moreover, several absolute gravity measurements were performed with FG5 to monitor the long term stability of the observatory which is set up on a 1270 m high karstic plateau. In this study, we compare the time series obtained from absolute gravity measurements, from combined LAGEOS-1 and -2 SLR solutions, and from GPS weekly solution for the vertical positioning, from 1998 to 2003. The geophysical annual vertical signal presents a magnitude of 5-6 mm. The absolute gravity signal presents amplitude of several Gal. Results are compared with different mass loading models. In particular, hydrologic and atmospheric loading models are applied, taking into account the local particularities of the OCA, which is on a plateau. We study the influence of the Mediterranean closed sea, which could act as a NIBO (Non-Inverted Barometer Ocean) where all pressure effects are fully transmitted to the Earth. In the global models, and particularly for the atmospheric delay correction, the proximity of the sea (south) and of the Alps (north) is investigated. The annual vertical signal observed in SLR and GPS vertical positioning time series is mainly explained by continental scale hydrological mass loading. The seasonal gravity signal coming from regional effects is isolated from local effects such as hydrology variations in the karst. Indeed, the difference between space geodetic positioning and gravity variations gives some indications concerning the hydrological variations and particularly the ground water table located under the OCA observatory. Acknowledgements: E. Calais, J.P. Boy, M. Amalvict, B. Luck, J.J. Walch, E. Gilli, M. Llubbes, P. Exertier