G11B-01
Temporal Mass Variations in Northern India as Observed by GRACE
Since successful launch of GRACE twin satellites, up to 70 monthly gravity field solutions have been released by several data processing centers. With a signal containing some seventy samples and sampling rate of one month, it is possible to significantly detect the periodicities between 2 and 35 months. In this study we applied the least- squares spectral method to analyze geoid changes over northern India and Nepal in the frequency domain, in order to determine the main existing temporal variations in that region. Results show some significant periodicities where the largest amplitude is for the annual cycle, as expected. A semi-annual amplitude and a signal with 2.5-year period are also included in the total mass changes. The results are presented and compared to hydrological models and further possible explanations for the mass variations (i.e. geophysical impacts or man-made effects) are given as well.
G11B-02
Singular spectral analysis of GRACE observations
Gravity Recovery and Climate Experiment (GRACE) launched 17.03.2002 gave the scientific community important information about the gravity field of the Earth and its changes under the influence of mass redistribution. These observations are especially valuable for the climate change investigations. Despite the all usefulness of the data, the nature of the experiment and its design imposes some limits on it. It is well known, that filtering is needed for GRACE measurements denoising and inversion. Finding the proper functional basis for signal representation is very important for successful filtering. Besides the spherical harmonic, wavelet, Slepian function basis, empirical orthogonal functions (EOF) are used by the scientific community (Rangelova E. et al., 2007; Wouters B. et al., 2007). EOF comes from the principal component analysis (PCA) of the correlation matrix, they represent the correlated patterns behavior in time and constitutes natural basis for the data. The extension of EOF analysis for the time series is singular spectral analysis (SSA) (Golyandina N. et al., 2002). In particular, for the varying fields multichannel SSA (MSSA) (Jollife I.T., 2002) allows to separate and represent spatial-temporal oscillatory components of the signal. This basis is also very good for filtering and prediction. We applied MSSA technique to monthly GRACE measurements of the gravity field of the Earth as in spatial as in spectral domain. The principal components were distinguished, related to different physical processes: water mass redistribution, ice melting, southern oscillation, etc. Annual oscillations, trend and noise were separated. The results were compared with models. Applied technique can be useful for GRACE data filtering, investigation of physical processes, which influence the gravity field distribution, for prediction of the gravity and climate changes. Comparison of the gravity field patterns with the sea level patterns evaluated by means of SSA from the altimeter measurements can help in separate investigation of the steric sea level rise and ocean mass change. The corrective smoothing is proposed by V.L. Panteleev for gravity measurements inversion. We show that this strategy can be successfully applied in the principal component basis instead of regularization.
G11B-03
Harmonic Analysis of Temporal Behavior of IGS Stations in Canada
The most recent approach of realizing the ITRF has been through rigorous combination of the position time
series from the Technique Centers (TC) together with daily Earth Orientation Parameters (EOP's) including
molar motion, universal time and length of day, with complete variance-covariance information. For this
ITRF2005 the technique centers provided space geodetic solutions from Very Long Baseline Interferometry
(VLBI), Satellite Laser Ranging (SLR), Global Positioning System (GPS) and Doppler Orbitography
Radiopositioning Integrated by Satellite (DORIS)].
The use of station position time series allows for the monitoring of non-linear station motion and
discontinuities as well as the temporal behavior of the origin and scale which are the frame physical
parameters. They also improve the consistency between ITRF and IERS EOP's. In view of this new approach of
the ITRF realization, the primary objective of this paper is to analyze the time series of GPS stations in Canada,
also known as the Canadian Active Control System (CACS), used as input to ITRF2005.
Using Least Squares Spectral Analysis we assess the temporal behavior from different geophysical
phenomena and assess their impact on the quality of the resulting ITRF2005 solutions. They include the
longtime variation due to datum shifts, technique errors (long-term orbit miss-modeling, long wavelength
multipath), as well as geophysical phenomena such as post-glacial rebound, Earth tides and non-tidal loading
effects.
http://www.iers.org/
G11B-04
Stochastic Characteristics of Modernized L2C Signal
Proper modeling of GPS system noise is essential for high-accuracy GPS positioning. This is especially important for modernized L2C signal, as existing models are based on analysis of limited data sets. In this contribution, we develop a new stochastic model for the L2C signal based on extensive analysis of long- duration data sets collected with the Trimble TR7 GNSS receivers using zero and short baselines. It is shown that the new model leads to better positioning solution and more realistic covariance matrix.
G11B-05
Satellite Orbit Perturbations in a Dusty Martian Atmosphere
We calculate the effect of Martian atmospheric dust on the orbital elements of a satellite. Dust storms that originate in the Martian surface may evolve into global storms in the atmosphere that can last for months and can affect low orbiter and lander missions. We model the dust as a velocity-square depended drag force acting on a satellite and we derive an appropriate disturbing potential function that accounts for the effect of dust on the orbit, using Lagrangean formulation. Furthermore we derive the dust drag coefficient Cd in order to calculate the orbital element time rates of the satellite. A first-order perturbation solution of Lagrange's planetary equations of motion using a 591km dust storm cloud that has a possible estimated density of 8.323´10-10 kg m-3 at an altitude of 100 km affects the semimajor axis of a 1000kg satellite by -0.036 m day-1. At the lower orbital altitude of h = 80 km a 591 km cloud with an estimated density of 3´10-9 kg m-3 results in a semimajor axis effect that is approximately 3.6 times higher than that at 100 km. The other orbital elements are also affected but to a lesser extend. In this paper we present detailed results on the effect of dust on all the orbital elements at various altitudes and how episodic dust may affect orbiter and lander missions.
G11B-06
Determination of Sea Surface Topography From Tide Gauge and Atmospheric Data
Tide gauges have been providing us with mean sea level information for more than a century nonetheless with
the inherent problems of their very limited spatial distribution, and the relative nature of the measurements with
respect to the moving crust. However, when combined with space geodetic techniques such as, GNSS and
satellite altimetry, tide gauges can provide the missing or weakly determined long-term component of the sea
level change as well as the mean sea level (MSL) near the coast. What is of importance in the definition of
modern reference height systems is the determination of the departure of the MSL from the equipotential
surface (geoid), known as sea surface topography (SST), which when determined, it will essentially provide a
"measurement" of the geoid (geoid=MSL-SST). The presence of a "topography" in the oceans (SST) is
explained by the dynamics of the oceans, which are primarily driven by atmospheric changes.
In this paper, we use very long records from tide gauges and meteorological stations from four Canadian
stations namely, Inukjuak (Quebec), Churchill (Manitoba), Victoria (British Columbia), and Argentia
(Newfoundland) to determine: a) sea level change, and b) SST, using a zero-frequency response method. Raw
hourly tide gauge data were obtained from the Marine Environmental Data Services (MEDS) of the Department
of Fisheries and Oceans (DFO) and the raw hourly atmospheric data (pressure, temperature and wind speed
and direction) at the same stations were obtained from Environment Canada.
For each station, we decimate all series using a low-pass filter with cut-off frequency of 0.03cph (8.33ìHz) to
eliminate most of the tidal signals and other high frequency content. Subsequently, we produce weighted least-
squares spectra of the filtered series by simultaneously suppressing long-term tidal constituents. The sea
level trend is obtained from the spectral analysis of the tide gauge data and the observed crustal uplift rates
determined by GPS observations. Similarly, the spectral analyses provide four atmospheric spectra namely for
pressure, temperature, and two wind speeds (one normal and the other parallel to the shore). We identify
unique common peaks, between the water level and each of the other four atmospheric spectra by using
appropriate combinations of product spectra. Amplitudes of common peaks at various frequencies are
determined and then used in a response analysis. Determination of the sea level response to the atmospheric
driving forces at zero frequency provides the permanent responses, and thus the SST. The preliminary results
we obtained from this independent approach for the sea level change are: Inukjuak +4.4 (+/- 2.7) mm/y;
Churchill +2.5 (+/- 0.7) mm/y; Victoria 0.0 (+/- 0.1) mm/y; Argentia +2.0 (+/- 3.1) mm/y, and for the SST: Inukjuak -
15cm; Churchill -18cm; Victoria -15cm; and Argentia -9cm. Comparisons of our results with those from other
methods will be discussed.
http://www.agu.org
G11B-07
The Absolute Gravimeter FG5 - Adjustment and Residual Data Evaluation
G11B-08
Crustal Deformation in western Sichuan Region and Implications for the May 12, 2008 Ms8.0 Earthquake
We have established 65 GPS stations and derive a detailed horizontal deformation distribution in western Sichuan region in conjunction with previous measurements at additional 90 stations. The resulting velocity field shows that strike-slip fault system here plays a significant role in adjusting crustal movements (in direction and magnitude) along the eastern borderland of the Tibetan Plateau. We find Xianshuihe fault is essentially creeping in its southeasternmost portion with left-lateral slip rate of 8.4±2.8mm/yr. We report sim5.6mm/yr left lateral slip rate across Anninghe and Daliangshan fault. Mabian fault slips left-laterally at the rate of 2.5±1.4mm/yr with nearly null fault-normal motion. Deformation across Litang and Yunongxi fault appear insignificant at present time. Our results reveal recently ruptured Longmenshan fault has both right-lateral slip rate and compressional rate close to 2.0±1.0mm/yr, implying that the Longmenshan fault indeed absorbs some northeastern motion of Tibetan Plateau and likely causes strain accumulation there, which leads to the breakout of the May 12, 2008 Ms8.0 earthquake. We also find several domains with markedly localized strain rates, which are associated with individual structures.
G11B-09
Inversion analysis of slip distribution of the 2008 Iwate-Miyagi Nairiku earthquake: Very high stress-drop or a conjugate fault slip?
On 14 June 2008, the Iwate-Miyagi Nairiku earthquake struck northeast Japan, where active seismicity has been observed under east-west compressional stress fields. The magnitude and hypocenter depth of the earthquake are reported as Mj 7.2 and 8 km, respectively. The earthquake is considered to have occurred on a west-dipping reverse fault with a roughly north-south strike. The earthquake caused significant surface displacements, which were detected by PALSAR, a Synthetic Aperture Radar (SAR) onboard the Japanese ALOS satellite. Several pairs of PALSAR images from six different paths are available to measure the coseismic displacements. Interferometric SAR (InSAR) is useful to obtain crustal displacements in the region where coseismic displacement is not so large (less than 1 m), whereas range and azimuth offsets provide displacement measurements up to a few meters on the whole processed area. We inverted the obtained displacement data to estimate slip distribution on the fault. Since the precise location and direction of the fault are not well known, the inverse problem is nonlinear. Following the method of Fukahata and Wright (2008), we resolved the weak non-linearity based on Akaike's Bayesian Information Criterion. We first estimated slip distribution by assuming a pure dip slip. The optimal fault geometry was estimated at dip 26 and strike 203 degrees. The maximum slip is more than 8 m and most slips concentrate at shallow depths (less than 4 km). The azimuth offset data suggest non-negligible right lateral slip components, so we next estimated slip distribution without fixing the rake angle. Again, a large slip area with the maximum slip of about 8 m in the shallow depth was obtained. Such slip models contradict with our existing common sense; our results indicate that the released strain is more than 10 to the power of -3. Range and azimuth offsets computed from SAR images obtained from both ascending and descending orbits appear to be more consistent with a conjugate fault slip, which contributes to lower the stress drop possibly to a level typical to this kind of earthquakes.
G11B-10
Determining Velocity Field and Strain Accumulation Throughout the Eastern North Anatolian Fault Zone
Serving as a bridge between Europe and Asia, Turkey is located at the intersection of these two continents. And North Anatolian Fault Zone (NAFZ) of Turkey is situated at a point where two of the Earth's tectonic plates meet, the Eurasian and Anatolian plates. The 1200 km long NAFZ runs along the northern part of Turkey, from Karliova in the east to the Gulf of Saros in the west connecting the East Anatolian compressional regime to the Aegean extensional regime. Therefore, eastern Turkey, the study area where NAFZ and East Anatolian Fault Zone (EAFZ) meets, is a complicated combination of active plate boundaries. Monitoring tectonic movements of this region is one of the most challenging activities of geodesy and geophysics. This region is capable of generating major earthquakes in every 3-4 years and needs to be monitored. In this study, we use Global Positioning System (GPS) velocities and solutions of earthquake focal mechanisms to determine a velocity field and obtain strain rates in the eastern part of NAFZ. We follow the method developed by Haines and Holt in 1993 to estimate strain rates and velocity model to carry out this study. Two sets of data are used: the first is geodetic velocities from the deformation network covered an area of 350 x 200 km square with 14 GPS stations, which provides significantly more data than ever for the region. The second is focal mechanism solutions of the earthquakes which occurred in the region since 1938. We obtain the results from the inversion of GPS data with constraints from seismicity. The results show that, strain is accumulating in the area, which may cause left-lateral release. It is expected that one or more of the major faults in the region will rupture and trigger an earthquake with a magnitude of at least 7.
G11B-11
Triangulation Based 3D Laser Imaging for Fracture Orientation Analysis
Laser imaging has recently been identified as a potential tool for rock mass characterization. This contribution focuses on the application of triangulation based, short-range laser imaging to determine fracture orientation and surface texture. This technology measures the distance to the target by triangulating the projected and reflected laser beams, and also records the reflection intensity. In this study, we acquired 3D laser images of rock faces using the Laser Camera System (LCS), a portable instrument developed by Neptec Design Group (Ottawa, Canada). The LCS uses an infrared laser beam and is immune to the lighting conditions. The maximum image resolution is 1024 x 1024 volumetric image elements. Depth resolution is 0.5 mm at 5 m. An above ground field trial was conducted at a blocky road cut with well defined joint sets (Kingston, Ontario). An underground field trial was conducted at the Inco 175 Ore body (Sudbury, Ontario) where images were acquired in the dark and the joint set features were more subtle. At each site, from a distance of 3 m away from the rock face, a grid of six images (approximately 1.6 m by 1.6 m) was acquired at maximum resolution with 20% overlap between adjacent images. This corresponds to a density of 40 image elements per square centimeter. Polyworks, a high density 3D visualization software tool, was used to align and merge the images into a single digital triangular mesh. The conventional method of determining fracture orientations is by manual measurement using a compass. In order to be accepted as a substitute for this method, the LCS should be capable of performing at least to the capabilities of manual measurements. To compare fracture orientation estimates derived from the 3D laser images to manual measurements, 160 inclinometer readings were taken at the above ground site. Three prominent joint sets (strike/dip: 236/09, 321/89, 325/01) were identified by plotting the joint poles on a stereonet. Underground, two main joint sets (strike/dip: 060/00, 114/86) were identified from 49 manual inclinometer measurements A stereonet of joint poles from the 3D laser data was generated using the commercial software Split-FX. Joint sets were identified successfully and their orientations correlated well with the hand measurements. However, Split-Fx overlays a simply 2D grid of equal-sized triangles onto the 3D surface and requires significant user input. In a more automated approach, we have developed a MATLAB script which directly imports the Polyworks 3D triangular mesh. A typical mesh is composed of over 1 million triangles of variable sizes: smooth regions are represented by large triangles, whereas rough surfaces are captured by several smaller triangles. Using the triangle vertices, the script computes the strike and dip of each triangle. This approach opens possibilities for statistical analysis of a large population of fracture orientation estimates, including surface texture. The methodology will be used to evaluate both synthetic and field data.