Carpathians Geodynamics Project

Does the upper mantle seismicity of the Romanian Vrancea Zone represent the last gasp of subduction of an embayment of the Tethys Ocean? Or is it actually a unique example of active continental lithosphere delamination? With Co-Principle Investigators Victor Mocanu, Laurentiu Munteanu, Mircea Radulian, Mihaela Popa, Klaus Bonjer, graduate student Mirela Dardac, and former UF undergraduate student Teresa Garcia, we are seeking to use the Vrancea Zone seismicity to characterize the upper mantle thermal state and current flow pattern in the Carpathian Arc region. The frequent seismicity from depths of 70-200 km beneath Earth's surface are well recorded at seismic stations in the Transylvanian Basin, the Eastern and Southern Carpathians, Dobrogea, the Moesian Platlform, and the Moldavian and Scythian portions of the East European cratonic platform. Seismic attenuation and anisotropy can help us resolve the question of the origin of the Vrancea seismicity, and the late Cenozoic geodynamics of the Transylvanian Basin, the Carpathian Arc, and the adjacent stable Platforms.

The following figures are from our publication:

Russo, R.M., V. Mocanu, M. Radulian, M. Popa, and K. Bonjer, Seismic Attenuation in the Carpathian Bend Zone and Surroundings, Earth and Planetary Science Letters, vol. 237, pp. 695-709, 2005.

Figure 1, above, shows the geography and topography of the Romanian Carpathians region. The concentration of red stars are the locations of intermediate depth Vrancea zone earthquakes. The blue-outlined fields are outcrop areas of mostly andesitic volcanic rocks erupted during the late Cenozoic, persumably as terranes in the Transylvanian Basin converged with the stable East Europen and Moesian Platforms. The heavy red line and dashed red line show geologists' best estimates of the suture line between Transylvania terranes and Platforms, and an extremal estimation (dashed line) of the suture. Note that the intermediate depth earthquakes - normally associated with subduction of lithosphere - lie SE of even the extremal potential suture.

Figure 2. Map of major regional structures, including active or recently active surface faults. Seismic stations of the German-Romanian K-2 Network, operated by the University of Karlsruhe and the Romanian National Institute of Earth Physics (NIEP) are shown as purple triangles. In the first phase of our study, we use recordings of Vrancea earthquakes at these stations to determine seismic attenuation and anisotropy in the study region. Note, however, that only one station (OZU) lies within the Transylvanian Basin.

Figure 3. Vrancea zone seismicity recorded and located by NIEP, plotted in a 3-D simulating perspective. View is from the NW looking SE. Topography plotted on box bottom as an aid to geographic recognition. Note that the Vrancea zone earthquakes define an oddly-shaped body with steep or nearly vertical plunge. "Shadows" of the seismicity are projected as black circles onto the eastern and southern sides of the box to clarify the body's shape.

Figure 4. Seismograms of Vrancea zone earthquake recorded at station VRI, almost immediately above the Vrancea zone (see Fig. 2). P and S waves are clear the window about each phase used in the iterative Q_S procedure shown as gray shaded areas about P and S. Also shown is pre-signal noise window used to estimate noise spectra.

Figure 5. P and S spectra for event shown in Figure 4. P spectrum measured from the vertical component, S spectrum from the transverse component. Note peaks of P and S energy do not coincide, and in order to avoid positive slopes, we limit the frequency band of calculation to the cross-over frequency.

Figure 6. Seismograms of Vrancea zone earthquake shown in Figure 4, detailing the two portions of each phase used in the iterative Q_S routine. P and S windows are each divided into two parts. The first portion of each window is always included in the spectral amplitude calculations. The second portion of the window is divided into 20 equal portions and sequentially added to the time series for P and S, respectively, before taking the Fourier transform. For each of the 400 (20x20) time series thus formed, the S-to-P spectral ratio is calculated and Q_S is determined. The mean Q_S and standard deviation are thus determined.

Figure 7. Each of the 400 spectra determined as described in Figure 6 is retained and summed to all its predecessors. This provides a composite or 'stacked' spectrum for both P (top left) and S (top right), filling holes in the spectra and suppressing noise (uncorrelated) as the time windows are lengthened. Noise spectra are shown as dotted lines. The spectral ratio of the normalized composite spectra is then determined and Q_S is calculated for the composite spectral ratio (bottom).

Figure 8. Ray paths from Vrancea events to stations on the East European, Scythian, and eastern Moesian Platforms, color coded according to observed Q_S for each event-station pair. Color code: white: S blocked; red: Q_S = 100-200; orange: Q_S = 200-250; yellow: Q_S = 250-300; green: Q_S = 300-350; cyan: Q_S = 350-400; blue: Q_S = 400-1000. Stations are magenta triangles and events are red stars. Topography, stations, and events shown projected on bottom of block.

Figure 9. Ray paths from Vrancea events to stations in the Carpathians and Transylvanian Basin, color coded according to observed Q_S for each event-station pair. Color code: white: S blocked; red: Q_S = 100-200; orange: Q_S = 200-250; yellow: Q_S = 250-300; green: Q_S = 300-350; cyan: Q_S = 350-400; blue: Q_S = 400-1000. Stations are magenta triangles and events are red stars. Topography, stations, and events shown projected on bottom of block.

Figure 10. Ray paths from Vrancea events to station LUC, in the eastern MoesianPlatform, color coded according to observed Q_S for each event-station pair. Color code: white: S blocked; red: Q_S = 100-200; orange: Q_S = 200-250; yellow: Q_S = 250-300; green: Q_S = 300-350; cyan: Q_S = 350-400; blue: Q_S = 400-1000. Stations are magenta triangles and events are red stars. Topography, stations, and events shown projected on bottom of block.

Figure 11. Comparison of Q_S observed at three stations lying along a line extending from the back-arc (OZU) across the Carpathians to the presumed forearc and foreland. Color-coded ray paths projected to surface. Note Q_S to OZU, in the Transylvanian Basin is almost all low (red and orange). At station GRE, results are strongly mixed, with apparently lower Q_S values from events in the southwestern portion of the Vrancea zone. At LUC, Q_S shows a pattern similar to that at GRE but less pronounced in Q_S variability. Heavy black line: NW-SE line of section for Figure 12. Color code: white: S blocked; red: Q_S = 100-200; orange: Q_S = 200-250; yellow: Q_S = 250-300; green: Q_S = 300-350; cyan: Q_S = 350-400; blue: Q_S = 400-1000. Stations are magenta triangles and events are red stars.

Figure 12a. Color-coded ray paths to three stations shown in Figure 11 projected onto NW-SE cross section. Tectonic units adapted from [Girbacea and Frisch, Geology, vol. 26, pp. 611-614, 1998], assuming a delamination horizon (heavy black-white dashed line) at 70 km, consistent with mantle xenolith compoisition in Persani basalts. High-Q paths to LUC and GRE cross preumably low-attenuation continental mantle lithosphere and crust. Not all the low Q_S rays to station OZU can be explained by this model, since many do not travel through highly attenuating asthenosphere. Either attenuation occurs at shallow depths beneath OZU, or the model should be modified.

Figure 12b. One possible modification to delamination model that would make it consistent with our results: raise the delamination horizon (heavy black-white dashed line) to a shallower depth so low-Q paths to OZU cross a significant thickness of asthenosphere.

Seismic Attenuation and Anisotropy in the Carpathians Region

Visit the UF-UB Carpathians Project Web Site

The University of Florida-University of Bucharest Project is sponsored by the Geophysics Program and International Programs of the U.S. National Science Foundation, and the University of Bucharest. We also gratefully acknowledge support from IRIS PASSCAL.



About Me GLY 2010 GLY 6932 CV Useful Links Home R. M. Russo Assistant Professor 223 Williamson Hall phone: 2-6766 rrusso@ufl.edu Office hours MWF: 11:40-12:40 or by appointment	Seismic Attenuation and Anisotropy in the Carpathians Region Does the upper mantle seismicity of the Romanian Vrancea Zone represent the last gasp of subduction of an embayment of the Tethys Ocean? Or is it actually a unique example of active continental lithosphere delamination? With Co-Principle Investigators Victor Mocanu, Laurentiu Munteanu, Mircea Radulian, Mihaela Popa, Klaus Bonjer, graduate student Mirela Dardac, and former UF undergraduate student Teresa Garcia, we are seeking to use the Vrancea Zone seismicity to characterize the upper mantle thermal state and current flow pattern in the Carpathian Arc region. The frequent seismicity from depths of 70-200 km beneath Earth's surface are well recorded at seismic stations in the Transylvanian Basin, the Eastern and Southern Carpathians, Dobrogea, the Moesian Platlform, and the Moldavian and Scythian portions of the East European cratonic platform. Seismic attenuation and anisotropy can help us resolve the question of the origin of the Vrancea seismicity, and the late Cenozoic geodynamics of the Transylvanian Basin, the Carpathian Arc, and the adjacent stable Platforms. The following figures are from our publication: Russo, R.M., V. Mocanu, M. Radulian, M. Popa, and K. Bonjer, Seismic Attenuation in the Carpathian Bend Zone and Surroundings, Earth and Planetary Science Letters, vol. 237, pp. 695-709, 2005. Figure 1, above, shows the geography and topography of the Romanian Carpathians region. The concentration of red stars are the locations of intermediate depth Vrancea zone earthquakes. The blue-outlined fields are outcrop areas of mostly andesitic volcanic rocks erupted during the late Cenozoic, persumably as terranes in the Transylvanian Basin converged with the stable East Europen and Moesian Platforms. The heavy red line and dashed red line show geologists' best estimates of the suture line between Transylvania terranes and Platforms, and an extremal estimation (dashed line) of the suture. Note that the intermediate depth earthquakes - normally associated with subduction of lithosphere - lie SE of even the extremal potential suture. Figure 2. Map of major regional structures, including active or recently active surface faults. Seismic stations of the German-Romanian K-2 Network, operated by the University of Karlsruhe and the Romanian National Institute of Earth Physics (NIEP) are shown as purple triangles. In the first phase of our study, we use recordings of Vrancea earthquakes at these stations to determine seismic attenuation and anisotropy in the study region. Note, however, that only one station (OZU) lies within the Transylvanian Basin. Figure 3. Vrancea zone seismicity recorded and located by NIEP, plotted in a 3-D simulating perspective. View is from the NW looking SE. Topography plotted on box bottom as an aid to geographic recognition. Note that the Vrancea zone earthquakes define an oddly-shaped body with steep or nearly vertical plunge. "Shadows" of the seismicity are projected as black circles onto the eastern and southern sides of the box to clarify the body's shape. Figure 4. Seismograms of Vrancea zone earthquake recorded at station VRI, almost immediately above the Vrancea zone (see Fig. 2). P and S waves are clear the window about each phase used in the iterative Q_S procedure shown as gray shaded areas about P and S. Also shown is pre-signal noise window used to estimate noise spectra. Figure 5. P and S spectra for event shown in Figure 4. P spectrum measured from the vertical component, S spectrum from the transverse component. Note peaks of P and S energy do not coincide, and in order to avoid positive slopes, we limit the frequency band of calculation to the cross-over frequency. Figure 6. Seismograms of Vrancea zone earthquake shown in Figure 4, detailing the two portions of each phase used in the iterative Q_S routine. P and S windows are each divided into two parts. The first portion of each window is always included in the spectral amplitude calculations. The second portion of the window is divided into 20 equal portions and sequentially added to the time series for P and S, respectively, before taking the Fourier transform. For each of the 400 (20x20) time series thus formed, the S-to-P spectral ratio is calculated and Q_S is determined. The mean Q_S and standard deviation are thus determined. Figure 7. Each of the 400 spectra determined as described in Figure 6 is retained and summed to all its predecessors. This provides a composite or 'stacked' spectrum for both P (top left) and S (top right), filling holes in the spectra and suppressing noise (uncorrelated) as the time windows are lengthened. Noise spectra are shown as dotted lines. The spectral ratio of the normalized composite spectra is then determined and Q_S is calculated for the composite spectral ratio (bottom). Figure 8. Ray paths from Vrancea events to stations on the East European, Scythian, and eastern Moesian Platforms, color coded according to observed Q_S for each event-station pair. Color code: white: S blocked; red: Q_S = 100-200; orange: Q_S = 200-250; yellow: Q_S = 250-300; green: Q_S = 300-350; cyan: Q_S = 350-400; blue: Q_S = 400-1000. Stations are magenta triangles and events are red stars. Topography, stations, and events shown projected on bottom of block. Figure 9. Ray paths from Vrancea events to stations in the Carpathians and Transylvanian Basin, color coded according to observed Q_S for each event-station pair. Color code: white: S blocked; red: Q_S = 100-200; orange: Q_S = 200-250; yellow: Q_S = 250-300; green: Q_S = 300-350; cyan: Q_S = 350-400; blue: Q_S = 400-1000. Stations are magenta triangles and events are red stars. Topography, stations, and events shown projected on bottom of block. Figure 10. Ray paths from Vrancea events to station LUC, in the eastern MoesianPlatform, color coded according to observed Q_S for each event-station pair. Color code: white: S blocked; red: Q_S = 100-200; orange: Q_S = 200-250; yellow: Q_S = 250-300; green: Q_S = 300-350; cyan: Q_S = 350-400; blue: Q_S = 400-1000. Stations are magenta triangles and events are red stars. Topography, stations, and events shown projected on bottom of block. Figure 11. Comparison of Q_S observed at three stations lying along a line extending from the back-arc (OZU) across the Carpathians to the presumed forearc and foreland. Color-coded ray paths projected to surface. Note Q_S to OZU, in the Transylvanian Basin is almost all low (red and orange). At station GRE, results are strongly mixed, with apparently lower Q_S values from events in the southwestern portion of the Vrancea zone. At LUC, Q_S shows a pattern similar to that at GRE but less pronounced in Q_S variability. Heavy black line: NW-SE line of section for Figure 12. Color code: white: S blocked; red: Q_S = 100-200; orange: Q_S = 200-250; yellow: Q_S = 250-300; green: Q_S = 300-350; cyan: Q_S = 350-400; blue: Q_S = 400-1000. Stations are magenta triangles and events are red stars. Figure 12a. Color-coded ray paths to three stations shown in Figure 11 projected onto NW-SE cross section. Tectonic units adapted from [Girbacea and Frisch, Geology, vol. 26, pp. 611-614, 1998], assuming a delamination horizon (heavy black-white dashed line) at 70 km, consistent with mantle xenolith compoisition in Persani basalts. High-Q paths to LUC and GRE cross preumably low-attenuation continental mantle lithosphere and crust. Not all the low Q_S rays to station OZU can be explained by this model, since many do not travel through highly attenuating asthenosphere. Either attenuation occurs at shallow depths beneath OZU, or the model should be modified. Figure 12b. One possible modification to delamination model that would make it consistent with our results: raise the delamination horizon (heavy black-white dashed line) to a shallower depth so low-Q paths to OZU cross a significant thickness of asthenosphere. Visit the UF-UB Carpathians Project Web Site http://seismology.geology.ufl.edu/romania The University of Florida-University of Bucharest Project is sponsored by the Geophysics Program and International Programs of the U.S. National Science Foundation, and the University of Bucharest. We also gratefully acknowledge support from IRIS PASSCAL.