G13C-01
Rotational Motion of the Victoria Plate with Respect to Nubia and Somalia Deduced from Space-Geodetic Observations
The East African Rift, a major tectonic structure that extends from the Afar region to the Southwest Indian Ridge mainly forming the plate boundary between Nubia and Somalia, is split into two branches in the Kenyan region, bordering a third tectonic block, called Victoria block. An angular velocity for this block was been recently proposed (Calais et al., 2006). However, it was based on earthquake slip vector mainly since only one GPS-derived motion was available in the limits of the block (Mbarara, Uganda). The lack of accurate GPS observations has prevented to obtain a more robust prediction of the relative motions of this block with respect to the Nubian and Somalian plates. In this work, we present a model of the present-day motions of the Victoria Block derived from the velocity field derived from a collection of GPS stations, both permanent- and campaign-type. The time-span of the observations is sufficiently long to derive reliable solutions for most of the stations in our network. Such model is compared with the most recent solutions based on the integration of sparse GPS data with earthquake slip vector. In order to properly interpret the motion solutions, the derived near-velocity field is constrained using a model of the present-day angular velocities of the stable part of Nubia and Somalia tectonic plates based on the ITRF2005 solution. The angular velocities are estimated using a network containing all available stations on Africa. We show that there is an anti-clockwise rotation of the Victoria block with respect to Nubia and Somalia plates, which in this region have a clearly relative extensional regime. The relative transverse motion of this block with respect to the neighbouring plates is about 2-3 mm/yr, which is approximately half of the differential motion between Nubia and Somalia in the area.
G13C-02
Present-day Kinematics of the East African Rift
The East African Rift (EAR), a ~5000 km-long series of seismically active structures that mark the divergent boundary between the Somalia and Nubia plates, is often cited as a modern archetype for rifting and continental breakup. Paradoxically, its current kinematics is the least well-known of all major plate boundaries, owing to its tremendous extent, difficult access, and lack of geodetic data. The existence of two "new" tectonic plates within the EAR (Victoria and Rovuma) between the main Nubian and Somalian plates has recently been proposed through combined analysis of earthquake slip vector and (sparse) space-geodetic data (Calais et al., 2006). A third, dominantly oceanic Lwandle plate was recently postulated by a re-analysis of the post 3.2 Ma spreading rates and transform-fault azimuths along the Southwest Indian Ridge (SWIR; Horner-Johnson et al., 2007). Here we use an updated geodetic solution -- a combination of all continuous GPS stations on the Nubia, Somalia, and Antarctic plates, campaign GPS measurements in Africa, and a global DORIS solution -- together with earthquake slip vectors in the EAR and transform azimuth and spreading rates along the SWIR. We show that the space geodetic data and the 3.2 Ma average oceanic data along the SWIR are consistent with each other at the 95% confidence level. The data support a kinematic model with (1) a present-day Nubia-Somalia Euler pole located to the SE of the southern tip of Africa, (2) the existence of three distinct microplates embedded in the EAR: Victoria, Rovuma, and Lwandle. Additional geodetic data in the EAR are now needed to confirm and further refine this model. References Calais, E., C. Hartnady, C. Ebinger, and J.M. Nocquet, Kinematics of the East African Rift from GPS and earthquake slip vector data, In: Yirgu, G., Ebinger, C.J. & Maguire, P.K.H. (eds) Structure and Evolution of the Rift Systems within the Afar volcanic province, Northeast Africa, Geological Society Special Publications, 259, p.9-22, 2006. Horner-Johnson, B., R.G. Gordon, D.F. Argus, Plate kinematic evidence for the existence of a distinct plate between the Nubian and Somalian plates along the Southwest Indian Ridge, Journal of Geophysical Research, 112, 2007
G13C-03
How rigid is a rigid plate? Geodetic constraint from the Kalahari craton, South Africa.
The motion and the rigidity of the Nubia plate provide critical constraint to the geodynamic of the surrounding plates. Unfortunately, the sparse distribution of geodetic continuous station across the plate does not allow to solve with high precision Eulerian pole and to test statistically for the rigidity. The presence of 3 separates cratons, rift valleys, and old deformation belts along the cratons" sutures, indicate that in geological time the plates have not been completely rigid. The amount of current deformation is very difficult to derive and from the gps observations has been constraint to be smaller than few mm/yr. Here, we present the velocity field for 42 stations of the continuous GPS network TRIGNET, a network covering the entire nation of South Africa with an average distance of 200 km. We present the velocity field for the period 2004- 2007 and the relative eulerian pole assuming rigid motion of the network. The distribution of these stations on the stable part of the Kalahari craton, allows computing a pole of rotation that can be compared with the rest of the stations within the Nubian plate. Preliminary results show that the entire network behaves as a rigid block with negligible average residual. Exceptions to this rigid motion are some of the costal stations and in the surrounding of Johannesburg. The analysis of the IGS stations on the rest of the plate using the eulerian pole from the TRIGNET network shows that the average residual are well within the errors indicating that the Nubia plate behave as a rigid block within our resolution and that the pole of rotation for the Kalahari craton can be used as eulerian pole for the entire Nubia plate. On the other hand, the non-random distribution of the azimuth of the residual does not exclude a possible CCW rotation of the craton with respect to the Nubia plate that at the stage cannot be distinguished from the noise.
G13C-04
Micro-deformation and Micro-seismicity in the Mak'arrasou Fault Zone in Afar, Kinematics Implications
The Mak'arrasou fault zone is formed of a multitude of N135-150°E-striking, right-stepping normal faults transferring the extension from the Asal Rift in the south to the Manda Inakir Rift, ~70 km to the north. Its overall NS direction and the regional extension in the N40°E direction (estimated by GPS) between the Arabia plate and the central part of the Afar depression, make the Mak'arrasou fault zone a left-lateral transfer zone. The seismicity is low in the south of the zone and inexistent or unknown in its northern part. We use the 10- year archive of RADARSAT InSAR data acquired from ascending and descending passes of the satellite to map two components of the surface velocity between 1997 and 2007. Beside the far-field movement across the Mak'arrasou fault zone, the most notable features of the deformation field during the 10-year period are localized areas of subsidence in the hanging wall of a normal fault in the southern part of the zone. The radar data analysis in time series shows that the largest subsidence of ~20mm occurred during two periods of slip on the fault in July and September 2001. The periods of ground movement correspond to swarms of small earthquakes in the vicinity of the fault, releasing a total moment of 8.75 1013 Nm over 45 days. We invert the integrated line of sight displacement fields from two stacks of 25 ascending and descending interferograms spanning the 3- month period to solve for the slip components on a small fault patch oriented N25°W, parallel to the local fault strike, and dipping 60° to the NE. The derived source mechanism is ~31 mm of slip with a rake of 82° from north, yielding a geodetic moment of 4.6 1015 Nm and consistent with a N46°E direction of extension. This direction is only ~6° east of the regional N40°E direction of extension, suggesting that the Mak'arrasou faults are accommodating the regional strain field with minor, if any, counter- clockwise rotation of the intervening blocks. This result is consistent with the absence of long-term rotation estimated using paleomagnetic observations in 1.8±0.4 Ma lava rocks in the southern part of the Mak'arrasou fault zone (Manighetti et al., 2001).
G13C-05
Twenty Years of GPS in Iceland: What Have we Learned About the Plate Boundary?
Iceland is located on the Mid-Atlantic Ridge, straddling the boundary between the North American and Eurasian plates. It is therefore one of the few places on Earth where direct observations of extensional processes related to plate spreading are possible. Since the first GPS experiment in 1986, most GPS campaigns have focused on limited parts of the plate boundary or individual volcanoes. Results from two nationwide GPS campaigns provide the first 3D velocity and deformation rate maps of the whole country during 1993--2004. The horizontal deformation appears to be accommodated by extension in a ~100 km wide region in North Iceland, whereas the zone of deformation in southern Iceland appears to be wider due to the overlapping western and eastern volcanic zones. The existence of a micro plate (the Hreppar block) has been suggested in the area between the two overlapping rift zones in southern Iceland. However, this has been difficult to confirm as the area is only about 100 km wide. In general, the observed relative horizontal motion agrees remarkably well with predictions from plate motion models, such as NUVEL-1A and REVEL. The horizontal velocities from 1993--2004 across the northern volcanic zone show that the velocity increase observed by GPS during 1987--1992, interpreted as a post-rifting signal following the 1975--1984 Krafla rifting episode, is no longer present. A significant rate of vertical uplift is observed in central Iceland during 1993--2004, with a maximum of the same order of magnitude as the rate of spreading (~20 mm/yr). The broad pattern of uplift can be explained by glacio-isostatic adjustment due to the thinning of Vatnajökull and three other large ice caps in Iceland since 1890. The residual vertical velocities indicate consistent areas of uplift or subsidence, mostly centered along the plate boundary. These observations and many of the previous GPS studies indicate that the elastic part of the lithosphere in Iceland is ~10 km thick, and lower crust/upper mantle viscosities are low (1-10 × 1018 Pa s). This suggests that transient deformation due to earthquakes and rifting events are a short lived phenomena (a decade) compared to the time interval between these events (hundred years). Presently, a significant portion of the strain built up by plate motion across the transform zone in south Iceland does not appear to be released seismically. It is not clear how the aseismic deformation in this region is accommodated in the upper crust.
G13C-06
Will Present day Glacier Retreat Cause Increased Volcanic Activity? Stress and its Effect on Magmatism induced by Glacier Retreat after the Little Ice Age at the Vatnajökull ice cap, Iceland
Global warming causes retreat of ice cap and ice sheets. Can melting glaciers trigger higher frequency of magmatic events? During the deglaciation of Iceland, at the Pleistocene-Holocene boundary, eruption rate is inferred to have been about 30-100 times its steady state. Increased decompressional mantle melting due to ice removal has been suggested as the main cause of the increase in melt production during deglaciation. A similar situation, on a smaller scale, exists in Iceland today. Since 1890 glaciers in Iceland have been continuously retreating, with the largest mass decrease occurring at the Vatnajökull ice cap. Ice melting is inducing crustal deformation around the ice cap with vertical velocities up to 25 mm/yr. We investigate the possible interactions between ongoing glacio-isostasy and an eventual change in melting rate or eruptive activity at volcanoes underneath Vatnajökull. A recent study of the glacio-isostatic deformation due to ice volume reduction at Vatnajökull since 1890 indicates a viscosity of the lower crust and upper mantle between 4- 10x1018 Pa s, assuming an elastic plate thickness of 10-20 km. Using these rheological parameters and a model of the ice retreat, we calculate the rate of change of pressure in the melting region. This pressure decrease is about 2x10-6 GPa/yr, which is an order of magnitude smaller than the rate of change of pressure decrease during deglaciation (about 1.9x10-5 GPa/yr). However, the current pressure decrease corresponds to moving the melting column upward by ~6 cm/yr. Stress changes also occur in the elastic layer, concentrating at the edges of the ice cap. Stretching of the crust, reducing pressure up to about -0.01 MPa/yr, takes place in the uppermost 5 km, while crustal compression of about 0.003 MPa/yr occurs between 5 and 10 km depth. These stresses are lower than those caused by stretching over the plate spreading region in Iceland (0.01-0.004 MPa/yr) but are likely to modulate the volcanic activity underneath Vatnajökull. The volcano most affected by these processes is the Bárdarbunga volcano, which is located at the ice cap edge. This volcanic system is inferred to have a deep magma chamber in the ductile layer probably connected to a shallow crustal magma chamber through a system of conduits. Glacio-isostatic stresses in the area may be such that both the deep and shallow magma chambers are subject to pressure decrease thus facilitating accommodation of residing magma, while compression of the conduits would prevent magma migration between the magma chambers. In this model current glacio-isostatic stress changes may reduce likelihood of volcanic activity at Bárdarbunga.
G13C-07
Global Plate Kinematics From GPS in Self-Consistent Reference Frame
Our study is based on the vectors of rotation of ten major lithospheric plates, that we estimated from continuous GPS observations at 192 globally distributed stations; 71 stations were selected as representing stable plate regions. The time series in the analysis span all days in the interval 1995-2007. In contrast to previous GPS plate models, we estimate relative plate rotations and plate-residual station velocities independently of any reference frame of the ITRF series. We analyze the drift of several ITRF origins relative to the center of plate rotation and show that the drift cannot be neglected in estimating plate motions, especially if the most recent ITRF2005 is used. The solution presented here is obtained under the constraint of no-net translation imposed on plate- residual velocities. Actually, we introduce the reference frame self-consistent with the plate kinematics, with the frame origin at the center of the sphere whose outer shell contains lithospheric plates. The model of the plate kinematics presented here addresses the problem debated since the start of the space geodesy, How big are disagreements between the current and geologic plate motions? We compare the vectors of relative plate rotations in our model with the published vectors from GPS and geologic models for several plate pairs discussed for over a decade. We also analyze the integrity of individual plates as evidenced by plate- residual station velocities. For seven largest plates, the rms value of plate-residual station velocities in stable plate interiors is 0.5-0.9 mm/yr; this value can be regarded as an upper bound on deviation of real plates from infinite stiffness.
G13C-08
ITRF2005 Plate Motion Model: A Sensitivity Analysis
The use of time series of station positions as input data for the construction of the ITRF2005 allows to control and eliminate station non-linear motion and all kind of discontinuities that would contaminate the long-term velocity estimates. We derived an absolute ITRF2005 Plate Motion Model using a velocity field of 152 sites located on rigid part of 15 tectonic plates. Given the 5 years more data and the greater number of sites involved, compared to what was used in case of ITRF2000 plate motion model, the angular velocities of the major plates appear to be better determined with the ITRF2005 data. However, this paper focuses on a comparative sensitivity analysis to a certain number of variants that impact the angular velocity estimates. The following variants are tested and the results of their comparisons are reported: number and distribution of sites per plate; solving for all plates motions simultaneously or separately; using full variance covariance matrix or diagonal terms only; assessing the impact of the Z-translation of 1.8 mm/yr between ITRF2000 and ITRF2005. Based on these tests we finally attempt to quantify the accuracy of the present day plate motion determination as derived from space geodesy measurements.