G43B-1192
Ground deformation detected by precise leveling just above the estimated pressure source in Asama volcano
Asama volcano Asama volcano erupted frequently in the first half of the 20th century, and almost stopped the eruption after then. However the volcano erupted in the top of volcano in September, 2004. Dike intrusion was estimated at the depth of 3km and 4km west from the Asama volcano from the ground deformation observed by GPS network around the volcano before and after the eruption (Murakami, 2005; Aoki et al., 2006). Precise leveling We started the precise leveling to the area just above the estimated dike intrusion from the mountain flank with a distance of 20 km in May, 2005 and repeated the leveling four times until May, 2007. Subsidence just above the estimated pressure source GPS measurements around the volcano shows the extension of the line lengths until July, 2005 and turn to the contraction after then. Leveling detected the uplift of 3mm on November, 2005, the subsidence on May, 2006, and the maximum subsidence of 10 mm just above the estimated dike intrusion on May 2007. Comparing the deformation with observed and calculated from source When we calculated the ground deformation by deflation of dike intrusion in semi-infinite elastic body using Okada model, it suggests the uplift just above the dike intrusion. As near field GPS network, established within 5 km from the summit after the eruption of September 2004 suggests a deflation of the volcano, one spherical deflation source is estimated in shallow area. Discussion of the refraction error of the precise leveling Refraction error of leveling is affected as an error of 30 mm in height measurements at the leveling route with the height deference of 800 m. The height deference of this leveling route is amounting to 1,000 m. As this leveling route is constructed in forest, refraction error is not so large.
G43B-1193
Geodetic and Modeling Constraints From Ongoing Uplift at Long Valley Caldera: Continued Episodic Dome Growth Observed in 2002-2003
Long Valley Caldera (LVC), a 17×32 km2 collapse crater in eastern California, sits on the eastern edge of the Sierra Nevada Mountains. Beginning around 1978, scientists observed signs of renewed unrest and developed a monitoring network including continuous seismic and geodetic observations. From that point, overall uplift of the central resurgent dome is estimated to be nearly 80 cm, from episodic events beginning around 1979(?), 1989, 1997 and 2002, along with interim periods of long-term uplift, inactivity, or slow deflation. While an understanding of the dynamics of this complicated time-history remains elusive, we focus here on the previously undescribed episode which occurred between 2002 and 2003. We use the continuous regionally filtered 3- component GPS data processed by the USGS, from which we remove the motion of the Sierra Nevada Block relative to North American plate (defined by Dixon et al., [2000]). Like deformation from the previous episodic uplift and interim periods, uplift is mostly radial, and can mostly be explained by a single source located just west of the central resurgent dome. The maximum uplift during the 2002-2003 episode is ~35 mm, about 1/3 the size but with a similar exponential shapes as the 1997-1998 episode. To better understand the dynamics of the system, we develop numerical finite-element models (FEM) to study the 2002-2003 inflation episode and determine the sensitivity of topography, layered rheology, and the presence of ring fractures on deformation. We find that while topographic effects are minimal, the choice of values for layered rheology and the existence of the south moat can contribute significantly to variations in observed deformation, and obscure the source location, particularly depth.
G43B-1194
Continuing subsidence and deformation of the Surtsey volcano, 1991 - 2002, Iceland
The Surtsey island was formed off the south coast of Iceland in a series of eruptions in 1963-1967. A submarine eruption was first detected on November 14, 1963, and the new island of Surtsey was built from the seafloor at 130 m depth. Two tephra cones formed above the sea level and when the magma conduit became isolated from the seawater the activity changed into lava effusion. A lava shield was produced as the lava flowed into the sea making a lava-fed delta. When the eruptions ceased the volume of Surtsey was about 0.8 km3 and with a surface of 2.65 km2. With crustal deformation observations on Surtsey, it is possible to follow the development of a pristine island. The vertical signal gives most information on the processes that are currently active. A levelling dataset, extending back to 1967, and the later GPS data give good opportunity to unravel the different processes. The GPS data improve the possibility to tie the vertical displacements to a reference frame outside the island and thus reduce the uncertainties in the absolute height determinations. Surtsey subsided rapidly during the first 10- 15 years and later on the rates decayed. The decaying rate was confirmed by GPS during 1992 to 2002. In the period 1992 - 2000 the rate was approximately 1 cm/yr, and for the 2000 - 2002 period approximately 0.5 cm/yr. The vertical displacement is not distributed uniformly over the island. The deformation processes currently active on Surtsey are: compaction of the volcanogenic material, slumping of the flanks of the island, the lithostatic load of the erupted material and compaction of the seabed sediments. Palagonitization of the tephra causes consolidation by growth of secondary minerals and thereby reduces the compaction. The amount of subsidence caused by the flexure of the lithosphere in this case is most dependent of the thickness of the elastic plate as it is the least constrained parameter. We estimate the most likely thickness to be 7-10 km. This gives a range of the total subsidence -0.53 m to -0.31 m from the load of 0.8 km3 erupted material. The compaction of the tephra and hyaloclastite is not uniform as the core of the island is palagonitized and consolidated while the flanks are still unconsolidated. The largest amount of subsidence is observed (15 cm in 11 years) along the side of the tuff cones where the lava overlays the delta. This vertical signal is non-uniformly distributed and is caused by submarine slope failure of the side of the island. The coastal erosion reduces the area of the island by one hectar per year, in the part of the island which consists of a lava-fed delta overlaid by sub-aerial lava flows. The palagonitized tuff cone is more resistant to the sea erosion as it virtually lacks open cracks. In the future only the palagonitized tuff cone will remain.
G43B-1195
Geodetic Characterization of Santorini Caldera From Continuous GPS Measurements
Santorini Caldera, in the southern Aegean, is part of a well developed, and very active volcanic system fueled by subduction along the Hellenic arc. The caldera is partially submerged, with only pieces of caldera wall, flanks, and central post-caldera lavas exposed above the sea level comprising a grouping of five small islands. The system had its most recent caldera-forming event around 1650 B.C. in a massive series of Plinean eruptions that expelled some 60 km3 of volcanic material, burying the previous island surface. The system remains active with ongoing smaller pyroclastic and phreatic eruptions, forming the central islets atop of the submerged caldera floor. In late-spring 2006, with UNAVCO field support and support form the Santorini Volcano Observatory, a network of two continuous GPS monuments spanning the caldera was established, and completion of a third monument is planned for this coming year. Additionally, 18 existing and new geodetic markers were first established with GPS across the 5-island group in 2006. These locations cover the caldera rim and flanks, and the central volcanic flows. Preliminary data from the two continuous GPS sites suggest that deformation across the caldera is currently minimal, and below the detection threshold for the 1.5 year continuous network. Through continuing analysis of the continuous network, along with additional campaign measurements, we hope to establish the temporal character and spatial extent of potential deformation in the volcanic complex, and determine if there exists any significant transient deformation associated with ongoing magma movement or edifice cooling. Monitoring such a rate over time may be useful for early hazard awareness and mitigation during regional volcanic crises.
G43B-1196
Intraplate Vertical Land Movements Constrained by Absolute Gravity Measurements
We have conducted repeated absolute gravity (AG) measurements across the tectonically active intraplate regions in Northwest Europe: the Ardenne and the Roer Graben. At most of the stations measurements were undertaken since 2000 and repeated twice a year. Our analysis of these measurements, performed in Belgium and Germany, show that at all stations except Jülich, there is no detectable gravity variation higher than 10 nm s-2 at the 95% confidence level. This is equivalent to vertical movements of 5 mm/yr. Although not yet significant, the observed rates do not contradict the subsidence predicted by glacial isostatic adjustment models and provide an upper limit on the possible uplift of the Ardennes. In Jülich, a gravity rate of change of 36.7 nm s-2/year equivalent to 18.4 mm/yr is due to anthropogenic subsidence. The amplitudes of the seasonal variations range from 18±0.8 nms-2 to 43±29 nms-2, depending on the location. These variations should have a negligible effect on the long-term trend, but at the Membach reference station, were a longer time series is available, differences in the rates observed since 1996 and 1999 indicate that long-term environmental effects may influence the inferred trend. The observed seasonal effects also demonstrate the repeatability of AG measurements. In Ostend, the AG time series agrees with tide gauge data, global mean sea level and altimeter measurements but disagrees with the CGPS. This study indicates that, even in difficult conditions, AG measurements repeated once a year can resolve vertical land movements at a few mm level after 5 years. This also confirms the need to measure for decades, using accurate and stable geodetic techniques like AG, in order to constrain slow deformation processes in an intraplate context.
G43B-1197
Elastic and viscoelastic crustal deformations in Greenland due to ice mass changes
We analyze data from ~10 continuous Global Positioning System (GPS) receivers and one tide gauge, all located along the edge of the Greenland ice sheet, to determine vertical uplift rates. We compare our results with predictions based on the ICE-5G deglaciation model of Peltier [2004]. Results from the GPS receiver at Kellyville and from the tide gauge at Nuuk, indicate that ICE-5G overestimates the subsidence rates at those locations by 1-2 mm/yr. Kellyville and Nuuk are located along the southwestern margin of the Greenland ice sheet, and the observed negative uplift rates are consistent with independent evidence that the ice margin along the southwestern edge readvanced during the last ~8 kyrs years to its current position. The ICE-5G glaciation- deglaciation history includes a readvance between the latitudes of 62 N and 72 N. The GPS measurements suggest the ICE-5G readvance may be too large. Our GPS results at Qaqortoq, located at the southern tip of Greenland, suggest a secular uplift rate of about -1 mm/yr, while ICE-5G predicts an uplift rate of 1 mm/yr. ICE-5G assumes no ice sheet readvance in south Greenland, including no readvance of the Qassimiut lobe (located at the southern tip of greenalnd) . The difference of 2 mm/yr can tentatively be explained as due to a ~60 km readvance of the Qassimiut lobe during the last ~3 kyrs. For the other GPS sites, the observed/predicted uplift rates are 3/-2 mm/yr at Kulusuk and indicate that ICE-5G does not exactly reproduce the correct rebound signal at those locations. The larger difference at Kulusuk, however, is probably due to the Earth's elastic response to ongoing changes in ice. The Kulusuk receiver is only ~90 km from the front of the Helheim glacier, where recent remote sensing observations have shown major periods of speedup between 2000 and 2005 [Howat et al., 2005] and thinning of the glacier by over 40 m from 2001 to 2003. We also analyze data from four continuous GPS receivers located between 0-150 km from the front of Greenland's Jakobshavn Isbr{\ae} Glacier. The GPS stations were established in bedrock to determine vertical crustal motion due to unloading of ice from Jakobshavn Isbrae. Results suggest a net loss of ~20 km3 of ice between May and August, 2006.
G43B-1198
Long term gravity change and rapid uplifting caused by glacial isostatic adjustment in southeastern Alaska
Glaciers at high latitudes are considered to be extremely sensitive to climate change and thus monitoring of glaciers is a clue to evaluate the future effect of global warming and the related phenomena. Ice mass changes also produce a time-variable surface load and give us useful data to investigate subsurface structure of the earth, especially to constrain the flow characteristics of the mantle. Larsen et al. (EPSL05) have extensively studied on vertical crustal movement in SE Alaska by means of raised shorelines, tide gauge measurements, and GPS to reveal the world's fastest glacial isostatic uplifting, which can be attributed to the response associated with glacier retreat. Displacement data, however, can only be used to constrain the sum of the elastic response to present-day ice melting (PDIM) and the viscoelastic one to past changes in ice. A Japan-US joint research project, ISEA (International geodetic research project in SouthEast Alaska), was initiated in 2005 to add new geodetic data sets and to refine the viscoelastic model derived by the previous studies. The outline of the project and some results are presented in this paper. In June, 2006, three kinds of field work were carried out. Absolute gravity (AG) surveys were performed at five sites in and around Glacier Bay using a Micro-g LaCoste absolute gravimeter, FG5#111 (Bilham and Sasagawa, EOS94). Gravity tide (GT) observation using a Scintrex's CG3M gravimeter was started in the campus of University of Alaska, Southeast (Sato et al., this meeting) to give precise corrections for the effect of ocean tide loading, which are the keys to increase the observation accuracy of AG and GPS. New continuous GPS (CGPS) sites were also established to examine not only the secular uplifting but the possible seasonal variation due to snow loading in the winter and ice loss in the summer. In December, 2006, another gravimeter, LaCoste and Romberg G-type #578, for GT observation with an electrostatic feedback circuit to increase stability of the scale factor (Harrison and Sato, JGR84) was installed at the same place of the CG3M gravimeter. In June, 2007, AG measurements were carried out at all sites of the 2006 survey and two other sites: Blanchard River in the Yukon Territory, Canada, where Sasagawa et al. (JGR89) performed AG measurements in 1987, and Fairbanks. We also conducted relative gravity survey using LaCoste-Romberg gravimeter G248 at some benchmarks, whose gravity values were compiled by Rice (1969). Some details of each fieldwork are presented below. We have derived long-term gravity changes at two sites, HNSG and BRM, where Sasagawa et al. (JGR89) conducted AG measurements in 1987. AG values obtained at HNSG in 2006 and 2007 by this project together with the value in 1987 (Sasagawa et al., JGR89) are used to estimate a regression line, which fits every observations very well. This suggests that the gravity at HNSG has decreased almost constantly for the last 20 years with the rate of -5.4ƒÝGal/yr, which is the world's largest as those caused by glacier retreat. Even though the AG values were obtained at BRM only twice, in 1987 and 2007, they also demonstrate the decreasing rate of - 3.9ƒÝGal/yr, which is consistent with the difference in the uplifting rate at two GPS sites near the AG sites obtained by Larsen et al. (EPSL05).
G43B-1199
Use of Coral Microatolls and a Tide Model to Measure Coseismic Vertical Deformation: Potential Utility and Common Mistakes
In the past few years, several great ( M>8) subduction megathrust ruptures have occurred beneath tropical seas and their fringing coral reefs. As predicted by elastic-dislocation theory, coastlines above the rupture patch rose, and adjacent regions subsided. Several investigators have used emerged or submerged coastal features to document land-level changes associated with these events. Unfortunately, when referencing these measurements to their high- or low-tide datums, some have overlooked the fact that both high and low tide levels can vary by more than a meter in some regions. In locations where tides are semidiurnal, a measured local high- tide value may be either the lower or higher high tide of the day, and a measured low-tide value may be either the higher or lower low tide. Furthermore, a measured high or low tide may be anywhere between the fortnightly spring and neap tides. Finally, even the elevations of spring and neap tides vary from month to month. One must know these variations to properly reference geological measurements. Some researchers have also made questionable assumptions about how the geological features they measured relate to tidal levels. As a result, published uplift or subsidence values in some studies may have errors of a meter or more, despite stated uncertainties of a few centimeters or less. A new approach, highlighted below, couples geological observations with a tide model to dramatically reduce uncertainties and produce more accurate estimates of uplift or subsidence. The upward growth of coral microatolls is controlled by low tide. Off the west coast of northern Sumatra, Porites microatolls' highest level of survival (HLS) is typically ~5 cm above annual low tide, but this is different for other genera and may be different in other regions. A comparison of pre- and post-earthquake HLS on a microatoll is the best method for documenting coseismic uplift; however, in cases where an entire reef was killed (and post-earthquake HLS cannot be found) or in cases of subsidence, a comparison of the pre-earthquake HLS to the post-earthquake annual low tide can provide an accurate and precise measurement of elevation change. Using a satellite- altimetry based tide model, which can determine tidal heights at any arbitrary time at any coastal location in the world, to properly tie any water level measurement into the tidal cycle (and hence to the annual low tide) for a site, it is possible to measure uplift or subsidence with realistic, demonstrable 2-sigma uncertainties of about a decimeter for sites with an overall tidal range of ~1m. While not as precise as continuous or even most campaign GPS measurements, this new approach allows for the determination of elevation changes with much greater spatial resolution than is normally achievable with GPS networks. With suitable adaptations, this approach can be used anywhere in the world.
G43B-1200
Vertical land displacements in Hokkaido, northern Japan
Pleistocene marine terraces along the north-eastern coast of Hokkaido island (Japan) are deformed as a result of the subduction of the Pacific Plate along the Kuril Trench. The morphology of Hokkaido, with terraces and coastal environments that extend orthogonal to the trench, provides a unique opportunity to test models of land movement over a variety of temporal scales. Current models suggest that the spatial pattern and rate of long- term deformation is constant over the last few interglacial cycles, and fossil shorelines from Hokkaido indicate that the region has experienced long-term uplift at a maximum rate of 76-90 cm/kyr around the Shiretoko Peninsula, 200 km from the Kuril Trench. In contrast, present-day land movements record subsidence along the eastern coast of Hokkaido, at a maximum rate of ca. 8 mm/yr (120 km from the trench). In this paper we compare vertical rates of deformation derived from different data sources over a range of time scales. We use continuous Global Positioning System observations and tide gauge measurements to reconstruct current and recent interseismic deformation around Hokkaido. Century-scale deformation over the last earthquake cycle (from c. AD 1650 to present) is evaluated with microfossil-based reconstructions of former sea-levels, whilst Holocene crustal movements are assessed by comparing calibrated sea-level index points with geophysical model predictions of relative sea-level. We apply forward elastic dislocation modelling using our estimates of crustal deformation on Hokkaido to investigate the characteristics of the Kuril subduction zone. Our work demonstrates the value of different types of palaeoseismic data for reconstructing spatial and temporal variations in the behaviour of the Kuril subduction zone and thus contribute to improved understand of seismic hazard in this part of the Pacific Ocean.
G43B-1201
Vertical component of the Earth's surface movement in the region of Central Europe (Czech Republic) from the results of satellite geodesy methods and their comparison with the results of repeated terrestrial geodetic methods.
Model EGG97 (European Gravimetric Geoid) was created for Europe well fitting even for the territory of the Czech Republic. Simultaneously the Geodynamic Network of the Czech Republic was founded for monitoring of recent dynamics of the Earth's surface. The sites of this network were observed by GPS(Navstar). At the end of the 19. century levelling observations were started, not too precise, but covering the whole territory of the Central Europe (Austro - Hungarian Empire) and observations were continuously precised. At the 20. century levelling network on the territory of Czech Republic was reobserved several times and the results of individual epochs were mutually compared. Long-term tendencies of the vertical component of the point positions at that dense network of levelling points were determined on the basis of these comparisons. Comparisons of the results of GPS observations of the sites of geodynamic network led (combined with the EGG97 geoid model) to conclusions that the differences of such determined heights with the levelled heights valid for height system of the Czech Republic contain, considering that the valid heights are more than 30 years old, very interesting information about the evolution of the height component of the positions of these points, which well correlate with comparable time intervals of the differences of repeated levellings carried out on the territory of the same extent. These demonstrable parallels of both more or less independent mothods of processing lead to the conclusions about vertical dynamics of the territory of the Czech Republic. Results are breafly completed by geological information, which explain some of the fundamental geodynamic tendencies of that territory. Geodetic results along with their geological interpretations can be the basis to work out the geodynamic model of the territory of Czech Republic.
G43B-1202
The annual horizontal and vertical movements of European Permanent Network stations in Central Europe
The contribution concerns about the horizontal and vertical movement of European Permanent Network (undermentioned EPN) stations in Central Europe. The movements are determined by the use of station coordinates from EPN weekly-combined solution. These coordinates are provided by GNSS technique and are for selected stations known since beginning of 1996 till today in week intervals. The stations with the longest time series and located in Central Europe are selected and their network is created. The time series analysis is carried out on the basis of station coordinates differences between two neighbour stations on one baseline. The adjustment of determined time changes in station differences on baselines is performed. The resultant movements in local coordination system (North, East, Up) are depicted and statistical analysis of results is performed.
G43B-1203
Seasonally Moving Bedrock Block at Metsähovi, in Finland
1. INTRODUCTION According to the monthly levelling observations during four years in Southern Finland at Metsähovi the bedrock block has been found out to move vertically 3.5 mm. The levelling accuracy in such small area is better than 0.1 mm. The vertical displacement has a strong correlation with the temperature of bedrock and the displacement is more than 10 times larger than caused by the thermal expansion alone (Lehmuskoski, 2006). Horizontal movements of this block have not been observed within the measuring accuracy of 0.5 mm. The rock type of moving block is a micmatite granite and its cleavage is typically orthogonal. Mainly the study area is without soil cap, and only a minor part of it is under a thin till layer. We have studied the area in many ways e.g. leveling bolts, GPR, the temperature of bedrock, and a core sample to find out a reason for a movement. So far we have not got an unambiguous answer, but we have eliminated several options. 2. INVESTIGATIONS By measuring the positions of leveling bolts we mapped roughly the dimensions of moving block. Moreover, the study area is measured by GPR with 2 m grid to locate all fractures with low dipping. According to these studies we have located a rock block with approximate dimensions of 30 × 40 × 1 m. Because of strong temperature dependency of the displacement and the fact that the seasonal temperature variation of bedrock is a near surface phenomenon we can conclude that only a shallow part of bedrock is moving. A groundwater table is deeper than the bottom of a moving block and the displacement doesn't have any correlation with a precipitation with time lag or without it. Thus, we can ignore a possibility that hydrostatic pressure could lift a block for instance after rains. As the surface of bedrock is at its lowest position at a winter time neither the frost can be the source of lifting. We drilled five shallow holes to locate exactly the bottom of moving block and to get signs of thermal expanding clays like Vermiculite. From one drill hole we got a core sample and we also image all holes to get better picture of fractures. None of these investigations or samples support that the source of observed displacement could be originated by clays. This is also supported by the fact that those clays do not occur in a granitic rock but prefer in a mafic rock (Deer, 1996). If thermal expanding cannot alone cause the observed movement, could some kind of lever system lift a block along with thermal expanding of rock? We know that vertical movement is the biggest in the center of block. This can be a sign from either the pending of edge fixed block results from the heavy horizontal pressure. We are going to install stretch slips to find out all tensions under which the surface of bedrock is. 3. CONCLUSIONS After intensive studies we have delineated well the shape of moving block. Still further investigations are needed to construct a proper 3D model of moving block. The origin of the displacement has strong temperature dependency and it must locate near to surface, because heat cannot penetrate deeply into rock in the short summer of Finland. As we have no evidence of the presence of any type of clays mechanical lever is the best explanation of movement. After the stretch slip test we know much more about the tensions of rock block. 4. REFERENCES Deer W. A., Howie R. A., and Zussman J., 1996. An Introduction to the Rock-Forming Minerals. Longman, China, pp. 696. Lehmuskoski P., Rouhiainen P., Saaranen V., Takalo M., and H., Virtanen, 2006. Seasonal Change of the Bedrock Elevation at the Metsähovi Levelling Test Field. Nordic Journal of Geodesy. Vol. 3, 1, 58 - 68.
G43B-1204
Earth tides observed by gravimeter and GPS in Juneau, Southeastern Alaska
We have analyzed the gravity data obtained at the Egan Library of University of Alaska, Southeast (UAS) and the GPS data obtained at a PBO site in Juneau, and we compared the obtained tidal amplitudes and phases with those estimated from the predicted tides including both effects of the body tide and ocean tide. To improve the accuracy in prediction, we also computed the regional ocean tide model in Southeastern Alaska (SE-AK). Our comparison results suggest; (1) By taking into account the ocean tide effect including the regional ocean tide model, the amplitude differences between the observation and the predicted body tide is remarkably reduced for both the gravity and displacement tides (e.g. for the M2 constituent, 8.04 microGal to 0.02 microGal (1 microGal=1E-8 m/s/s), and 2.43 cm to the order of 0.01 cm for the vertical displacement), (2) The PPP method (Zumberge et al., 1997) used to extract the tidal signals from the original GPS time series works well to recover the tidal signals. Although the GPS analysis results still contain the noises which may be considered to be the meteorological effects, we may conclude that the GPS observation surely detects the tidal signals with the sub cm accuracy or better for some of the tidal constituents. The viscoelastic effect in gravity tides is estimated to be the order of 0.05 ?Gal for the M2 constituent in Juneau. The magnitude is equivalent to the analysis error for our gravity data. Increasing the accuracy of calibration of the gravimeter and the regional ocean model in SE-AK is needed to constrain the Earthfs viscoelastic response to the tidal force tightly.
G43B-1205
Accurate modeling of regional oceanic tides for detiding gravimeter data at southeastern Alaska
We use a regional barotropic ocean model and investigate oceanic tides in southeastern Alaska characterized by complicated fjords to achieve the accurate prediction of the oceanic tides. The regional ocean model is forced by major 8 component tides of FES2004 (Lyard et al., 2006) at the computational boundaries. The simulated semidiurnal tides at coasts facing the open oceans are in good agreement with the observation. At the inside region of the Alaska Panhandle (AP), the simulated semidiurnal tides change in phase with the observation with their amplitude larger than the observational amplitude. Especially the amplitude of M2 tide is simulated by up to 40-50 cm larger than the observational value. The semidiurnal tides are amplified at the inside AP. Diurnal tides are realistically simulated at all coastal stations without amplification at the inside AP. The main fjord at the inside AP is assumed to be a one-dimensional channel. The period of the surface seiche of the channel is basically determined by the basin geometries, and is roughly estimated to be 9 hours. Therefore the sea level at the inside AP probably resonates with the semidiurnal tides. The gain of the semidiurnal tides seems strongly sensitive to uncertainties of the seabed topography of the inside AP.
G43B-1206
Using GPS-Derived Vertical Movements For Studies of Crustal Loading on Fennoscandian Shield
The surface of the Earth is constantly being deformed by temporally and spatially varying loads of atmospheric, oceanic and hydrological masses. These deformations are detectable in the height component of GPS time series. We have studied the loading effect using GPS-derived vertical movements for three stations on Fennoscandian shield, Metsähovi, Onsala and Bor\rmås. Three loading factors that are not yet routinely implemented in the GPS calculations have been assessed, namely loading caused by air pressure variation, by non-tidal sea level changes and by changes in the continental water reservoirs. The effect of continental water was studied using two global soil moisture models. Air pressure data was taken from a local barometer in Metsähovi and from the numerical weather model HIRLAM (High Resolution Limited Area Model) for the stations Onsala and Bor\rmås. Sea level data was obtained from the nearest tide gauges. Multilinear regression analysis was used for air pressure and sea level changes. The loading factors are compared with the vertical motion observed by GPS for a time period of 4 years 8 months. The results show that the amplitude of the annual signal varies from 4.0 mm at Metsähovi to 1.2 mm at Borås. The correction for the soil moisture seems to remove part of the annual signal. When all the factors are considered, the variance of the GPS time series diminishes by 40% for Metsähovi, 24% for Onsala and 15% for Bor\rmås. The regression coefficients for the air pressure were -0.32 mm/hPa for Metsähovi, -0.27 mm/hPa for Onsala and -0.25 mm/hPa for Bor\rmås. The coefficients for the sea level height were -6 mm/m, -8 mm/m and -5 mm/m, respectively. The reduction of the loading phenomena dimishes the scatter of the GPS time series. The loading corrections could be used for several purposes, e.g. for the studies of other error sources, to create stable time series and also to correct for campaign measurements, which take place in varying weather conditions.
G43B-1207
Response of Hydrated Crust to Tidal Loading: Theoretical Predictions for Comparison With GPS-inferred Displacements.
In recent years, deformation of the Earth surface due to ocean-tide loading has been measured by both ground- and space-based techniques in several coastal regions. In particular, displacements inferred from the Global Positioning System (GPS) have been reported for regions of high tidal ranges (e. g. British Isles; Bay of Fundy). As the database of such measurements lengthens, the displacement amplitudes and phase lags of the main tidal constituents improve, thereby providing better constraints on possible ocean tide loading displacement (OTLD) models. Available OTLD models, however, are all based on the assumption that the response of the solid Earth at tidal frequencies is perfectly elastic. This assumption is questionable at the regional scale, especially where the earth structure is predominately deformed by the higher harmonics in the loading function. Under such conditions, ground-based geophysical constraints indicate an earth structure which is heterogeneous, fractured, dissipative and often fluid-saturated. A data-constrained structural model that reflects these complexities is presented; it is used to predict the anelastic earth-response to ocean tide loading at the primary semi-diurnal period (12.25 h). Central to this model is an absorption mechanism by which elastic energy is lost to heat via viscous flow of interstitial nanofluids. The scaling of this microscopic mechanism to observational scales is done using an effective-medium scheme and the theory of viscoelasticity. The anelastic response is quantified by a quality factor (Q) which, in turn, quantifies the phase lag of response behind loading. Predictions from such modeling are compared to calculations from currently used OTLD models, and to recent GPS-derived tidal displacements. Thus, demands on the precision required of GPS data for significantly improving OTLD models can be assessed.
G43B-1208
Yearly gravity variations of superconducting gravimeter (GWR C039) at Ny-Å lesund compared to GPS data and hydrological model
Gravity data from the superconducting gravimeter (GWR C039) at Ny-Å lesund, Svalbard, Norway, have been analysed for long periodic variations. Prior to the analysis the gravity data was filtered and cleaned, by removing e.g. earthquakes and spikes. Any drift in the SCG data was removed using available absolute gravity measurements. Yearly variations in the gravity signal are visible after removal of short periodic tides, i.e. tides with period of approximately one month and shorter. The gravity station is co-located with 2 IGS-GPS stations. The GPS recievers are located within 5 m from the superconducting gravimeter. The variations in gravity data may be due to changes in the snow/water level. We investigated this effect by comparing to GPS data and global hydrological models. The effect of changes in the level of sediment deposits and glaciers are evaluated.
G43B-1209
Atmospheric loading coefficients determined from homogeneously reprocessed long-term GPS and VLBI position time series
The paper uses homogeneously reprocessed VLBI (OCCAM6.1e, at DGFI) and GPS (Bernese, at GFZ) long-term position time series to determine atmospheric loading coefficients. As the VLBI and GPS data analysis is done using fully homogenized models (for troposphere etc), and fully reprocesses all data, the analysis results are expected to give a clear insight into the level of similarity one can reach with both techniques in terms of station position time series (concerning annual signals etc.). A comparison of the VLBI- and GPS-derived atmospheric loading coefficients with each other and a comparison of the signal derived from these coefficients with existing time series of atmospheric loading signals will allow an assessment of how well they may be interpreted.
G43B-1210
Modeling and Observation of Time-Variable GPS Site Positions
Our aim is to investigate loading consequences on the time-variable GPS site positions of hundred stations around the world during the period 2003-2006. We model the three dimensional site displacements using a Love number formalism to describe the elastic deformation of a spherical Earth's model submitted to atmospheric, oceanic and hydrological loadings. We produce site positions time series using the GPS analysis software GAMIT/GLOBK with/without inserting a combination of loading models and study their impact on 3D site positions. We first of all compare the variability of modeled and observed sites positions (without loading in the GAMIT Software). We secondly study the variance reduction in the GPS sites positions provided by the loading process. We conclude that the seasonal variability of sites displacements is quite well explained by our model in several geographic regions, e.g. at mid-latitudes in the northern hemisphere, while it is much less understood at other locations, e. g. near coastal areas.
G43B-1211
GHYRAF (Gravity and HYdrology in AFrica): an Experiment to Validate GRACE in Africa From the Sahara to the Equatorial Monsoon Zone
We present an experiment to validate GRACE observations with multi-disciplinary data (gravity, geodesy, hydrology, and meteorology) based on the observation of seasonal changes in water storage in Africa. We foresee to concentrate on two areas: the desert zone in the Sahara that provides a null test with almost no hydrological changes, and the equatorial monsoon band that provides on the contrary a large rainfall signal. Our proposal includes two types of ground-based gravity measurements. First we will perform a repeated survey with absolute gravimeters (AG) on a North-South profile during a 2-3 year time span (2008-2010) to assess the large soil moisture changes as predicted by existing hydrological models. Second we plan to establish a superconducting gravimeter (SG) to act as a continuously monitored base station in a region of large soil moisture changes. In addition, continuous geodetic GPS measurements will be made along the profile to assess the vertical deformation which is not seen by GRACE. In-situ measurements of hydrological parameters at each station will assist us in modelling local gravity effects. We will point out the main difficulties arising in the ground validation of satellite-derived gravity observations. On one side, we will estimate the role played by the vertical motion of the measurement point located on the Earth's moving surface that alters gravity as measured by AG or SG on the contrary to satellites. On the other side, we will pay attention to the length scales involved in hydrological processes which are differently retrieved whether gravity is measured at the ground or by satellite. Examples using recent models of continental hydrology such as LadWorld or GLDAS will be given.
G43B-1212
GPS observation of compaction or expansion of the Perth basin aquifer system
Solutions from reprocessing over ten years (1997-2006) of continuous GPS (CGPS) data from the Australia region using absolute GPS satellite antenna and receiver antenna phase correction models in the latest IGS05 reference frame (the IGS realisation of the ITRF2005 reference frame) have made estimation of vertical velocity at a resolution of 1mm/yr practicable. The accuracy of CGPS vertical velocity can reliably detect a small vertical land motion caused by groundwater level change in response to prolonged periods of groundwater pumping and recharge. Three CGPS sites, namely HAL1 (Hillary), PERT (Perth) and YAR2 (Yarragadee) located in the Perth Basin in Western Australia, record subsidence of 5mm/yr (in HIL1 and PERT) and uplift of 2mm/yr (in YAR2). These changes are all statistically significant using different short memory and long memory time series noise models. The key source for the changes is correlated with groundwater levels. In the Perth metropolitan area, potentiometric head in the confined aquifers has declined significantly over the past decade in response to high groundwater extraction rates due to drought conditions. The GPS vertical velocity in HIL1 and PERT reflects elastic compaction of the aquifer system. In agricultural areas north of Perth, such as typified by the location of YAR2 CGPS station, groundwater levels have risen by as much as 0.4m/yr during recent decades. This is due to increased recharge through replaced perennial native vegetation by annual pasture and crops. Accordingly, upward GPS vertical velocity in YAR2 may reflect vertical expansion of strata in the semi-confined aquifer system. In addition, the significant annual signal in YAR2 is strongly correlated to the seasonal rainfall change in the region.
G43B-1213
Vertical displacement by absolute and relative gravimetry in Central Taiwan: comparison with GPS, leveling and INSAR results
Yunlin and Changhua, located in central Taiwan, are two agricultural counties where significant surface subsidence has occurred over the past decade. The major culprit is pumping of ground water. Measures to stop the subsidence have been taken, but the subsidence continues. To detect the extent and magnitude of recent subsidence here, we use a FG5 absolute gravimeter, and GRAVITON-EG and SCINTREX-5 relative gravimeters to determine gravity changes and consequently subsidence rates. The gravimetric results show varying subsidence rates with the maximum reaching 10 cm/year. Hydrological effects introduce uncertainties to the gravimetric results and require dedicated modeling of ground water, soil moisture and other temporal gravity variations. Ocean tide loadings and seismic effects of earthquakes and ocean dynamics also add to the uncertainties. The gravimetric results are compared with the results from GPS, leveling and PSINSAR. Good agreement between these results is found, despite some large, scattering discrepancies.
G43B-1214
Probing the Uppermost Mantle Rheology Using Surface Deformation Associated with the Lake Mead Load Fluctuations
Water level fluctuations in the Lake Mead (Basin and Range, USA) act as a time varying load on the lithosphere. The monitoring of the surface deformation induced by the water load is performed by InSAR using ERS-1, ERS-2 and Envisat radar acquisitions between 1992 and 2007. The comparison between load and deformation could bring constraints on the uppermost mantle viscosity structure, provided a sufficient accuracy of InSAR measurements. Interferograms from ERS-1 and ERS-2 SAR acquisitions were previously calculated and inverted to retrieve the temporal and spatial subsidence around lake Mead between 1992 and 2001. A peak to peak subsidence of 1.5 cm is recorded between 1995 and 1999 and corresponds to 10m of lake level increase. It was shown in a previous study that the ground motion evolution can be closely associated with the load/unload of lake Mead water level fluctuations (Cavalié et al., J. Geophys. Res, 2007). The ground motion amplitude, evolution, and pattern could be best explained by a visco-elastic rebound of the lithosphere- asthenosphere, with a 30 km thick elastic plate overlying a 1018 Pa.s asthenosphere. In order to provide further constraints on the lithosphere and asthenosphere rheology, it was found necessary to extend the monitoring of surface deformation up to 2007. We need, in particular, to record the uplift that is associated with the drastic lake level fall (25 m) from 2000 to 2004. This should theoretically allow to measure not-in-phase ground motions with respect to loading, that were not yet clearly visible in the 1992-2001 time series. To achieve that goal, we use Envisat data from 2003 to 2007, and a few additional ERS-2 data from 2004 to 2007, acquired with a "reasonable" Doppler centroid frequency. The link between the ERS 1992-2001 and the Envisat 2003-2007 time series is obtained through a few cross-platform ERS/Envisat interferograms and a few ERS-2 interferograms covering the data gap in 2002-2003. We shortly describe the mitigation of atmospheric artefacts. More than 400 ERS/ERS, ERS/Envisat and Envisat/Envisat- interferograms, performed with the JPL ROI-PAC software, are then included in an inversion to obtain maps of the ground subsidence from 1992 to 2007. We give the elastic thickness and asthenosphere viscosity that allow the best adjustment between the model and the spatial and temporal behavior of the estimated displacement. Although extended to 2007, the ERS-Envisat displacement time series accuracy is still marginally acceptable to place strong constraints on a complete asthenosphere/lithosphere layered mechanical behavior.