1
Velocity Structure of the Iran Region Using Seismic and Gravity Observations E. Bergman U Colorado at Boulder H. Zhang U of Science and Technology of China E.M. Syracuse, M. Maceira, W.S. Phillips, M.L. Begnaud Los Alamos National Laboratory Iran and Surrounding Area Data Work in progress The views expressed here do not necessarily reflect the view of the United States Government, the United States Department of Energy, or the Los Alamos National Laboratory. This work is sponsored by the Department of Energy, National Nuclear Security Administration, and Defense Nuclear Nonproliferations R&D and Los Alamos National Laboratory’s Laboratory Directed Research and Development. The recent tectonics of Iran are dominated by continental collision between Arabia and Iran and the subduction of the Indian Ocean plate beneath the Makran. These plate interactions produce a dominant N20E compressional strain across much of the country. On a broad scale, deformation within Iran is localized within the Zagros, Alborz, and Kopet-Dag mountainous regions, although moderate and deadly seismic activity occurs along the boundaries of the relatively stable Central Iran, Lut, and South Caspian blocks. The region contains some relatively sharp geologic transitions, notably the continental-ocean boundary and the transition from the Iranian plateau to the Russian platform northeast of Iran, as well as com- plexities associated with the Caspian and Black Sea basins. With the exception of the Late- Proterozoic age Arabian Shield, much of the focus area consists of rela- tively young, Phanerozoic terranes accreted onto the southern Eurasian plate during the closing of the Tethys Ocean. 40° 40° 45° 45° ° 0 5 50° 55° 55° 60° 60° 65° 65° 70° 70° 25° 25° 30° 30° 35° 35° 40° 40° 45° 45° 40° 40° 45° 45° ° 0 5 50° 55° 55° 60° 60° 65° 65° 70° 70° 25° 25° 30° 30° 35° 35° 40° 40° 45° 45° 1851 1945 1765 6.5 6.0 5.0 8.0 7.5 7.0 USGS - NEIC 1970-2012 Alborz Mountains Caucasus Mts Makran Subduction Zone Makran Arabia Eurasia South Caspian Turan Platform Helmand Block Kopet Dag Lut Block Persian Gulf Mesopotamian Foredeep North Caspian Zagros Mountains Volcanic Center map courtesy C. Ammon 40° 40° 50° 50° 60° 60° 70° 70° 20° 20° 30° 30° 40° 40° -150 -100 -50 0 50 100 150 Free-air Gravity Anomaly (mgal) Gravity observations Seismic body wave travel times P- and S-arrivals from 3750 earthquakes occurring between 1973 and 2013 are included. Only earth- quakes that satisfy the condition of having a second- ary azimuthal gap < 180°, calculated over the entire range of epicentral distances are included. The sec- ondary azimuthal gap is the largest azimuth gap that is filled by a single station. Focal depths are color- coded. Gravity data are extracted from the global Earth Gravitational Model EGM2008 - pub- licly released by the U.S. National Geospatial-Intelligence Agency. Free-air gravity anomalies for are shown here. Data are filtered prior to the joint inversion to remove long-wavelength features. Surface wave dispersion curves Rayleigh wave group velicites are mea- sured from vertical-component data. 2D frequency-dependent tomographic in- versions result in velocity maps for fre- quencies between 10-34 Hz. These maps are sampled at the grid nodes used for the joint inversion, and these resulting disper- sion curves are used as input. Results Based on checkerboard tests with different anomaly sizes and velocity perturbations, we conclude that we can robustly resolve 5 layers in the crust and upper mantle with a horizontal grid parameterizrion of 1 x 1 degrees. Regularization parameters and relative weighting of the different geophysical datasets play an important role in obtaining accurate inversion results. In an attempt to minimize subjectivity while choosing the appropriate parameters, we performed multiple inver- sion runs with different combinations of damping and smoothing. Determing the appropriate relative weight- ing of gravity, surface waves, and body waves is cur- rently in progress. Preferred compressional velocity (Vp) model (upper left) and shear velocity (Vs) model (near left) for the body wave-only inversion using the optimal set of damping and smoothing, as determined from the tradeoff analysis. Warm colors indicate slower velocities and cold colors indicate faster than average velocities. The color scale is different for each depth slice. 25˚ 30˚ 35˚ 40˚ 25˚ 30˚ 35˚ 40˚ 25˚ 30˚ 35˚ 40˚ 25˚ 30˚ 35˚ 40˚ 25˚ 30˚ 35˚ 40˚ 25˚ 30˚ 35˚ 40˚ 40˚ 45˚ 50˚ 55˚ 60˚ 25˚ 30˚ 35˚ 40˚ 40˚ 45˚ 50˚ 55˚ 60˚ 25˚ 30˚ 35˚ 40˚ 25˚ 30˚ 35˚ 40˚ 25˚ 30˚ 35˚ 40˚ −5 0 5 25˚ 30˚ 35˚ 40˚ 25˚ 30˚ 35˚ 40˚ −5 0 5 25˚ 30˚ 35˚ 40˚ 25˚ 30˚ 35˚ 40˚ −5 0 5 40˚ 45˚ 50˚ 55˚ 60˚ 25˚ 30˚ 35˚ 40˚ 40˚ 45˚ 50˚ 55˚ 60˚ 25˚ 30˚ 35˚ 40˚ −5 0 5 25˚ 30˚ 35˚ 40˚ 20 km 25˚ 30˚ 35˚ 40˚ 25˚ 30˚ 35˚ 40˚ 40 km 25˚ 30˚ 35˚ 40˚ 25˚ 30˚ 35˚ 40˚ 60 km 25˚ 30˚ 35˚ 40˚ 40˚ 45˚ 50˚ 55˚ 60˚ 25˚ 30˚ 35˚ 40˚ 90 km 40˚ 45˚ 50˚ 55˚ 60˚ 25˚ 30˚ 35˚ 40˚ 25˚ 30˚ 35˚ 40˚ 25˚ 30˚ 35˚ 40˚ −5 0 5 dVp (%) 25˚ 30˚ 35˚ 40˚ 25˚ 30˚ 35˚ 40˚ −5 0 5 25˚ 30˚ 35˚ 40˚ 25˚ 30˚ 35˚ 40˚ −5 0 5 40˚ 45˚ 50˚ 55˚ 60˚ 25˚ 30˚ 35˚ 40˚ 40˚ 45˚ 50˚ 55˚ 60˚ 25˚ 30˚ 35˚ 40˚ −5 0 5 2 x 2 Input Recovered Input Recovered dVp (%) dVp (%) dVp (%) dVp (%) dVp (%) dVp (%) 20 km 40 km 60 km 90 km 1 x 1 Vp model Checkerboard Tests Tradeoff analysis Vs model 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 V p model length (km/s) 800 1000 1200 1400 1600 P arrival misfit (ms) 0 4 8 12 16 20 24 28 V s model length (km/s) 1500 2000 2500 3000 S arrival misfit (ms) damping 50 100 150 200 250 300 400 500 smoothing 20 200 400 600 800 900 1000 1100 1200 1400 1600 initial misfit, 1D model optimal model · damping=150 · smoothing=1200 0 20 40 60 80 100 120 140 160 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 0 20 40 60 80 100 120 140 160 180 0 5 10 15 20 25 30 35 Can multiparameter tomography improve travel time predictions and source locations? Code optimization Adding surface waves & gravity Travel times only model used for re- location of 1892 events: (1) median displacement is 3.82 km (2) 90% level is 10.1 km Joint model used for relocation of 1892 events: (1) median displacement is 3.69 km (2) 90% level is 9.9 km The first-order goal of tomography is to derive improved models of Earth structure. The nature of “improvements” is not easy to judge, however. Here, we try to address the question: “Can multi-parameter tomography address crustal heterogeneities and areas of limited coverage, and improve travel time predictions?” The implication of this question is that improvements in theoretical travel times calculated from a model based on multi-parameter tomog- raphy could lead to improved location accuracy. We consider the extent to which relocating the events in the joint inversion moves them away from the starting locations. We present two versions of this test, one based on an in- version of travel time data only, and the second being a joint inversion of travel-time and gravity data. Ray tracing using the finite-difference method employed in our code JointTomoFDD is computationally expensive for a region the size of Iran. A new version of the code using a modified pseudo-bending ray-tracer is currently being implemented, which will allow a thor- ough and robust testing of regularization parameters and relative data weightings. In addition, recent improvements to the gravity component of the joint inversion will help better recover velocities within the crust. Dispersion curves resulting from recent surface wave analyses are currently being incorporated to the joint inversion. These are expected to greatly improve shear-wave velocity recovery and help constrain deeper portions of the velocity model, below depths sampled by body waves from the deepest regional seismicity. 35˚ 40˚ 45˚ 50˚ 55˚ 60˚ 65˚ 70˚ 20˚ 25˚ 30˚ 35˚ 40˚ 45˚ 2.2 2.6 3.0 3.4 V Rayleigh (km/s) 20 s Period 35˚ 40˚ 45˚ 50˚ 55˚ 60˚ 65˚ 70˚ 20˚ 25˚ 30˚ 35˚ 40˚ 45˚ 2.4 2.8 3.2 3.6 V Rayleigh (km/s) 30 s Period
