Search

Accurate seismic velocity models are a necessary parameter for seismic hazard analysis in high hazard regions. Therefore, we present the California-Nevada Adjoint Simulations (CANVAS) model, an adjoint waveform tomography model of California and Nevada that is optimized to fit waveforms. We began by testing different starting models for CANVAS to determine what effect starting model choice plays on final inversion results. Though studies have compared synthetic seismograms to compare starting models (e.g., Zhou et al., 2021), this is the first study to compare results after inversion. We chose CSEM_NA (Krischer et al., 2018), SPiRaL (Simmons et al., 2021), and WUS256 (Rodgers et al., 2022) as starting models and iterated each model to a minimum period of 20 seconds. We then compared the results of the final models using five comparison metrics. All five of the comparison metrics showed that the final models all resolved tectonic structure and fit observed data similarly, regardless of the structure or fit of the starting models. Therefore, we conclude that the choice of starting model has minimal effect on the final model results. Errors in source mechanisms can produce bad synthetic seismograms, which limit the data available to use in adjoint waveform tomography inversions. Many researchers use source mechanisms from the Global Centroid Moment Tensor (GCMT) catalogue. However, the GCMT catalogue has known issues with resolving shallow crustal earthquakes. Since shallow crustal earthquakes dominate CANVAS’s dataset, we inverted for source mechanisms using MTTime, a time-domain moment tensor inversion code. We calculated 3D Green’s functions using an intermediate version of CANVAS that was iterated to a minimum period of 20 seconds. We showed that the inverted source mechanisms greatly improved waveform fit compared to the GCMT solutions, particularly at distant stations. We also demonstrated that 3D Green’s functions have better and more consistent waveform fits co

The SPE-N is a series of chemical explosions intended to enhance our physical understanding and ability to quantitatively model seismic signals from explosions at the NNSS with the goal of improving our nuclear test monitoring capabilities (Brunish et al., 2010). The main objective is to develop comprehensive three-dimensional high fidelity modeling capabilities to enable a transition from the current empirically-based explosion monitoring approach, where accuracy hinges on the availability of calibration data, to a physics-based predictive approach, where seismic observables are correlated to both non-linear physical processes in the near source region, and wave propagation scattering away from the source. To address these issues, we are developing an end-to-end three-dimensional (3D) simulation methodology to model the data collected, including non-linear (shock) motions and linear anelastic motions in the solid earth. We seek to understand the partitioning of energy excited by underground explosions and its propagation in complex 3D geological structure to far-field seismic monitoring stations. The data obtained during the SPE series will be used to validate and calibrate these numerical modeling techniques. Although the entire wavefield will be modeled, particular focus is on the generation of shear waves., Published in the Proceedings of the 2012 Monitoring Research Review - Ground-Based Nuclear Explosion Monitoring Technologies, 18-20 September 2012, Albuquerque, NM. Volume II. Sponsored by the Air Force Research Laboratory (AFRL) and the National Nuclear Security Administration (NNSA). U.S. Government or Federal Rights License

The Middle East is a tectonically complex and seismically active region. The ability to accurately locate earthquakes and other seismic events in this region is complicated by tectonics, the uneven distribution of natural earthquakes and the fact that countries run separate national seismic networks without well-developed data-sharing agreements. We report here on a variety of scientific efforts to enhance knowledge of the lithospheric velocity structure in the Middle East, making use of data from national networks in Saudi Arabia, Oman, and Kuwait to improve seismic location accuracy for events throughout the region. Collaborative seismology engagements with research institutes in the Middle East have produced several important findings. Studies of regional seismic structure revealed that the Arabian Shield and the Arabian Platform have fundamental differences in velocity structure. The earth?s crust is relatively thicker than average in Kuwait and Iraq as a result of the Mesopotamian Foredeep, characterized by 8-10 km thick sediments. Ophiolites dominate the southeastern margin of the Arabian Peninsula, where the crust is thicker and seismic velocities are faster. In Kuwait seismic activity is intensified near oil fields, and seismic source parameters show that this activity comes from tectonic events. This work is informed by continuous or event-based regional seismic data obtained from 68 seismic stations from national networks, and this number is expected to grow in the future. A current effort is focused on using regional seismic data to develop better and more accurate seismic magnitude scales based on coda for stations in these national networks. Data and derived measurements from stations are integrated into lithospheric velocity and attenuation models to increase resolution, improve event location accuracy and source parameter determination, and advance tectonic understanding of the region., Published in the Proceedings of the 2012 Monitoring Research Review - Ground-Based Nuclear Explosion Monitoring Technologies, 18-20 September 2012, Albuquerque, NM. Volume I. Sponsored by the Air Force Research Laboratory (AFRL) and the National Nuclear Security Administration (NNSA). U.S. Government or Federal Rights License

This project involves research being performed to improve ground motion simulation capabilities for underground explosion monitoring. We are working along two thrusts: 1) we are coupling hydrodynamic (non-linear shock) and seismic (linear anelastic) wave propagation codes; and 2) we are investigating the effect of source emplacement conditions on ground motions. For both thrusts we are modeling explosion motions using GEODYN, an Eulerian hydrodynamic code developed at Lawrence Livermore National Laboratory (LLNL). This code includes many important features for modeling shock waves in geologic materials, including non-linear response (e.g., effects of porosity, tensile failure, yielding) and adaptive mesh refinement. However, numerical solution of the hydrodynamic response is computationally expensive due to non-linear constitutive behavior, especially when compared to elastic wave propagation solvers. To propagate ground motions from the non-linear explosive source region to far-field seismic stations we are using a one-way coupling strategy to pass motions from GEODYN to WPP (LLNL's anelastic finite difference code for seismic wave modeling). Motions computed by GEODYN and recorded on a dense grid span the ranges where motions become linear (elastic). These are saved, processed and passed to WPP where they are introduced as a boundary source and continue to propagate as elastic waves at lower numerical cost than with GEODYN. In the past year we have worked on several important details to gain confidence that coupled GEODYN-to-WPP simulations are accurate. This involved modifying GEODYN to accurately model elastic surface waves. This is challenging because GEODYN uses hydrodynamic rather than elastodynamic equations of motions for the continuum., Published in the Proceedings of the 2011 Monitoring Research Review - Ground-Based Nuclear Explosion Monitoring Technologies, 13-15 September 2011, Tucson, AZ. Volume I. Sponsored by the Air Force Research Laboratory (AFRL) and the National Nuclear Security Administration (NNSA). U.S. Government or Federal Rights License

The Source Physics Experiment (SPE-N) at the Nevada National Security Site (NNSS) is planned as a series of chemical explosions under a variety of emplacement conditions. The SPE-N goal is to improve our physical understanding and ability to model how explosion generate seismic waves, particularly S-waves. The first SPE-N explosion (SPE1) occurred in May 2011 and consisted of a 220 lb (100 kg) chemical explosion at a depth of 180 ft (55 m) in granite (Climax Stock). This paper examines the far-field seismic observations and a complementary paper addresses the near-field wave-motion (Antoun et al. [these Proceedings]). The shot was well-recorded by a assortment of over 150 instruments. The greatest density of instruments were located within 5 km of the source but some extended as far out as 20 km, and there are additional local and regional stations that are part of permanent networks. The majority of the SPE-N specific stations were installed as part of 5 radially oriented lines that emanate from the shot location. A variety of instruments were deployed, including high-frequency geophones accelerometers, broadband seismometers, and rotational sensors. Data recovery from the first shot was over 95%. A review of the SPE1 waveforms shows variations with azimuth in both frequency content and amplitudes. Some of the variations correlated closely with known lateral changes in geology but the nearest stations, all of which were located on granite, also showed significant variations. Preliminary modeling of the SPE1 data using a 3D finite difference code and a 3D velocity model showed similarities in both arrival times and overall features at low frequencies. Further modeling will be conducted with emphasis on understanding the generation of shear waves and comparison of data with the synthetics., Published in the Proceedings of the 2011 Monitoring Research Review - Ground-Based Nuclear Explosion Monitoring Technologies, 13-15 September 2011, Tucson, AZ. Volume I. Sponsored by the Air Force Research Laboratory (AFRL) and the National Nuclear Security Administration (NNSA). U.S. Government or Federal Rights License

The main goal of this project is to develop models of lithospheric velocity structure for Eurasia and the Middle East in order to improve capabilities within National Nuclear Security Administration (NNSA) labs to accurately predict travel times for local and regional phases, as well as travel-times for body waves at upper-mantle triplication distances. Velocity models of the lithosphere are key for accurately modeling not only travel times but also surface-wave dispersion velocities and full waveforms at regional (2 deg - 15 deg) and far-regional (15 deg - 25 deg) distances. The models are being developed following a two-step approach: first, one-dimensional (1D) velocity models for select broadband stations are obtained by jointly inverting P- and S-wave receiver functions and fundamental-mode group and phase dispersion velocities; second, regionalized velocity models are constructed by combining the 1D joint inversion models within regions used in the UNIFIED model and are validated through regional waveform modeling. We expect the velocity models will also help inform and strengthen ongoing and future efforts within the NNSA labs to develop three-dimensional (3D) velocity models for Eurasia and the Middle East, assist in obtaining model-based predictions where no empirical data are available, and improve event locations in regions with sparse network coverage. So far, we have obtained a total of 59 joint inversion models in Eurasia and the Middle East: 35 for Europe, 10 for the Middle East, and 14 for Asia. To develop these models, we have considered permanent broadband stations with open access and with waveforms archived at the Data Management Center (DMC) of the Incorporated Research Institutes for Seismology (IRIS). The station distribution is quite uneven among the UNIFIED regions, and the selected stations have been complemented with open stations archived at other data centers when available., Published in the Proceedings of the 2011 Monitoring Research Review - Ground-Based Nuclear Explosion Monitoring Technologies, 13-15 September 2011, Tucson, AZ. Volume I. Sponsored by the Air Force Research Laboratory (AFRL) and the National Nuclear Security Administration (NNSA). U.S. Government or Federal Rights License

We present the latest results on Lawrence Livermore National Laboratory's calibration efforts for seismic attenuation of regional body and surface waves that have application to many different areas of nuclear explosion monitoring. We have developed methods that use amplitude measurements of the direct regional phases (Pn, Pg, Sn Lg) to determine the attenuation structure of the lithosphere in Eurasia. The amplitudes are inverted simultaneously for attenuation parameters (Qp, Qs) of the crust and upper mantle, along with event source terms and station site terms. We are applying similar methodologies to coda amplitudes. Like direct waves, coda waves are subject to path-dependent variations in amplitudes. We see geographic similarities between the crustal shear-wave attenuation and the results from the coda attenuation. Calibration of coda in the Middle East and other areas is complicated by the fact that the dominant S-wave phase is either Sn or Lg depending on tectonic region, distance, and frequency. Over the past year, we have made great progress on the calibration of surface wave attenuation with the development of the Surface Wave Amplitude Processor (SWAP). With this tool, we are able to make surface wave amplitude measurements quickly, reliably, and consistently. We will be presenting a preliminary surface wave attenuation tomography of the Middle East. Regional attenuation models are directly applicable to event discrimination, such as high-frequency regional P/S discriminants (e.g., Pn/Lg, Pg/Lg, Pn/Sn) and longer period Ms:mb. Correcting the observed amplitudes for path-dependent variations reduces scatter in the earthquake population and increases separation from explosions. Better body-wave path corrections might even allow the extension of P/S discrimination to lower frequencies so long as true source differences between events exist at those frequencies., Published in the Proceedings of the 2011 Monitoring Research Review - Ground-Based Nuclear Explosion Monitoring Technologies, 13-15 September 2011, Tucson, AZ. Volume I. Sponsored by the Air Force Research Laboratory (AFRL) and the National Nuclear Security Administration (NNSA). U.S. Government or Federal Rights License

Improving current Middle East wave speed models with full waveforms required confidence in sources and recordings, along with a methodology to iteratively improve models and reduce the minimum period of the waveforms used. A large, well recorded subset of 201 events (1997 2007) was reinterpreted through a direct comparison between data and synthetics based upon a centroid moment tensor inversion. Initial evaluations were done using a 1D reference model at periods greater than 80 seconds and a more stringent evaluation was done for a three-dimensional (3D) model at periods of 25 seconds and longer. Final source reinterpretations within the 3D model define a source database and the initial starting point for tomography. Transitioning from a 1D to 3D wave speed model shows dramatic improvements when comparisons are done at shorter periods (25 s) and even at longer periods (80 s). Updates of the wave speed model were accomplished using adjoint tomography. Initial attempts to update the wave speed model were hampered by the strong anisotropy in the mantle causing an unavoidable mismatch between Rayleigh and Love waves when using an isotropic mantle parameterization., The original document contains color images. Published in the Proceedings 2011 Monitoring Research Review: Ground-Based Nuclear Explosion Monitoring Technologies Tuscon AZ v1 pp161-167. Prepared in collaboration with Princeton University, and Lawrence Livermore National Laboratory.

Advances in numerical methods for modeling seismic wave excitation and propagation and ever more powerful parallel computers are making it easier to simulate ground motions on the scale lengths (domain sizes) and frequencies (resolutions) of importance to nuclear explosion monitoring. The objective of this research is to develop and improve methods for seismic simulation in fully 3D earth models to improve nuclear explosion monitoring. Specifically, research is directed along three thrusts: modeling of shock-wave propagation with hydrodynamic methods; modeling of elastic propagation near shallow explosions and earthquakes, including the effect of 3D volumetric structure and free-surface topography; and regional broadband waveform modeling. This effort relies on numerical methods for 3D wave phenomena implemented in computer codes running on massively parallel computers. These capabilities allow us to investigate source and propagation phenomenology of explosion sources and understand waveform data from specific events of interest., Published in Proceedings of the 2010 Monitoring Research Review - Ground-Based Nuclear Explosion Monitoring Technologies, 21-23 September 2010, Orlando, FL. Volume I. Sponsored by the Air Force Research Laboratory (AFRL) and the National Nuclear Security Administration (NNSA). U.S. Government or Federal Rights License

It is well known that one-dimensional models do a poor job of predicting both regional amplitudes and travel-times over large and tectonically complicated regions. As a result regional discrimination methods (e.g., high-frequency P/S, Ms:mb) and magnitude estimates (e.g. coda magnitude) can perform poorly when applied over broad regions. The careful calibration of the earth's attenuation structure is critical to the universal application of event discrimination and yield estimation methods down to very small magnitudes. We have developed and are continuing to improve methods that use the direct amplitudes of the major regional phases (Pn, Pg, Sn, Lg) to determine the attenuation structure of the lithosphere in the Middle East and East Asia. The amplitudes are inverted simultaneously for attenuation parameters (Qp, Qs) of the crust and upper mantle, event source terms, and station site terms, which ensures that parameters, such as seismic moment and apparent stress, are consistent across each phase. We are using these models to correct observed amplitudes for path-dependent variations due to earth structure. Amplitude corrections can be demonstrated to improve high-frequency regional P/S discriminants (e.g., Pn/Lg, Pg/Lg, Pn/Sn) by reducing scatter in the earthquake population and increasing separation from explosions. Better path corrections allow the extension of discrimination to lower frequencies so long as true source differences between events exist at those frequencies. We are applying similar methodologies to coda amplitudes with the goal of improving magnitude and yield estimation. While coda waves average over large regions and have less variation than direct phases, they too are subject to path-dependent variations in amplitudes. Furthermore, they can be sensitive to changes in the dominant phase (e.g., Sn vs. Lg) over broad regions. We report on our efforts for the 2-D calibration of coda amplitudes in the Middle East., Published in Proceedings of the 2010 Monitoring Research Review - Ground-Based Nuclear Explosion Monitoring Technologies, 21-23 September 2010, Orlando, FL. Volume I. Sponsored by the Air Force Research Laboratory (AFRL) and the National Nuclear Security Administration (NNSA). U.S. Government or Federal Rights License

The main goal of this project is to develop regionalized models of lithospheric velocity structure for a wide variety of tectonic regions throughout Eurasia and the Middle East. We expect the regionalized models will improve the ability of the National Nuclear Security Administration (NNSA) labs to predict travel times for local and regional phases, such as Pg, Pn, Sn and Lg, as well as travel times for body waves at upper-mantle triplication distances in both seismic and aseismic regions. The models have been developed following a two-step approach: (i) first one-dimensional (1D) velocity models for select broadband stations are obtained by jointly inverting P- and S-wave receiver functions and fundamental-mode group and phase dispersion velocities, and (ii) regionalized velocity models are then constructed by combining the 1D joint inversion models within each tectonic region and validated through regional waveform modeling. The velocity models thus obtained will also help inform and strengthen ongoing and future efforts within the NNSA labs to develop 3D velocity models for Eurasia and the Middle East, and will assist in obtaining model-based predictions where no empirical data are available and for improving locations from sparse networks. During the first year of this project, we have developed 1D velocity models for 54 locations in Europe and 10 locations in the Middle East. Receiver functions were computed from teleseismic P- and S-waveforms recorded at open broadband stations and archived at the Data Management Center of the Incorporated Research Institutes for Seismology (IRIS-DMS), while dispersion velocities were obtained from an independent surface-wave tomography study for Eurasia and North Africa. Due to a combination of short recording time-windows and inefficient recording of teleseismic S-waves, some locations in Western Europe did not yield reliable S-wave receiver function estimates., Published in Proceedings of the 2010 Monitoring Research Review - Ground-Based Nuclear Explosion Monitoring Technologies, 21-23 September 2010, Orlando, FL. Volume I. p87-96. Sponsored by the Air Force Research Laboratory (AFRL) and the National Nuclear Security Administration (NNSA). U.S. Government or Federal Rights License

Efforts to update current wave speed models of the Middle East require a thoroughly tested database of sources and recordings. Recordings of seismic waves traversing the region from Tibet to the Red Sea will be the principal metric in guiding improvements to the current wave speed model. Precise characterizations of the earthquakes, specifically depths and faulting mechanisms, are essential to avoid mapping source errors into the refined wave speed model. Errors associated with the source are manifested in amplitude and phase changes. Source depths and paths near nodal planes are particularly error prone as small changes may severely affect the resulting wavefield. Once sources are quantified, regions requiring refinement will be highlighted using adjoint tomography methods based on spectral element simulations (Komatitsch and Tromp, 1999). An initial database of 250 regional Middle Eastern events from 1990-2007, was inverted for depth and focal mechanism using teleseismic arrivals (Kikuchi and Kanamori, 1982) and regional surface and body waves (Zhao and Helmberger, 1994). From this initial database, we reinterpreted a large, well-recorded subset of 201 events through a direct comparison between data and synthetics based upon a centroid moment tensor inversion (Liu et al., 2004). Evaluation was done using both a ID reference model (Dziewonski and Anderson, 1981) at periods greater than 80 seconds and a 3D model (Kustowski et al., 2008) at periods of 25 seconds and longer. The final source reinterpretations will be within the 3D model, as this is the initial starting point for the adjoint tomography. Transitioning from a ID to 3D wave speed model shows dramatic improvements when comparisons are done at shorter periods, e.g., 25 s. Synthetics from the ID model were created through mode summations while those from the 3D simulations were created using the spectral element method. To further assess errors in source depth and focal mechanism, comparisons between the thr, The original document contains color images. All DTIC reproductions will be in black and white. Published in the Proceedings of the 2009 Monitoring Research Review, Ground-Based Nuclear Explosion Monitoring Technologies, held in Tucson, AZ, on 21-23 September 2009, vI p194-200, 2009.

We continue exploring methodologies to improve earthquake-explosion discrimination using regional amplitude ratios such as P/S in a variety of frequency bands. Empirically, we demonstrate that such ratios separate explosions from earthquakes, using closely located pairs of earthquakes and explosions recorded on common, publicly available stations at test sites around the world (e.g., Nevada, Novaya Zemlya, Semipalatinsk, Lop Nor, India, Pakistan, and North Korea). We are also examining if there is any relationship between the observed P/S and the point source variability revealed by longer period full waveform modeling (e.g., Ford et al., 2008). For example, regional waveform modeling shows strong tectonic release from the May 1998 India test, in contrast with very little tectonic release in the October 2006 North Korea test, but the P/S discrimination behavior appears similar in both events using the limited regional data available. While regional amplitude ratios such as P/S can separate events in close proximity, it is also empirically well known that path effects can greatly distort observed amplitudes and make earthquakes appear very explosion like. Previously we have shown that the Magnitude Distance Amplitude Correction (MDAC) technique (Walter and Taylor, 2001) can account for simple 1-D attenuation and geometrical spreading corrections, as well as magnitude and site effects. However, in some regions, 1-D path corrections are a poor approximation, and we need to develop 2-D path corrections. Here we demonstrate a new 2-D attenuation tomography technique using the MDAC earthquake source model applied to a set of events and stations in both the Middle East and the Yellow Sea Korean Peninsula regions. We believe this new 2-D MDAC tomography has the potential to greatly improve earthquake-explosion discrimination, particularly in tectonically complex regions such as the Middle East., Published in the Proceedings of the Monitoring Research Review: Ground-Based Nuclear Explosion Monitoring Technologies (30th) held in Portsmouth, VA on 23-25 Sep 2008, p702-711, 2008. The original document contains color images.

The development of accurate numerical methods to simulate wave propagation in three-dimensional (3D) earth models and advances in computational power offer exciting possibilities for modeling the motions excited by underground nuclear explosions. This presentation will describe recent work to use new numerical techniques and parallel computing to model earthquakes and underground explosions to improve understanding of the wave excitation at the source and path-propagation effects. Firstly, we are using the spectral element method (SEM, SPECFEM3D code of Komatitsch and Tromp, 2002) to model earthquakes and explosions at regional distances using available 3D models. SPECFEM3D simulates an elastic wave propagation in fully 3D earth models in spherical geometry with the ability to account for free surface topography, anisotropy, ellipticity, rotation and gravity. Results show in many cases that 3D models are able to reproduce features of the observed seismograms that arise from path-propagation effects (e.g., enhanced surface wave dispersion, refraction, amplitude variations from focusing and defocusing, tangential component energy from isotropic sources). We are currently investigating the ability of different 3D models to predict path-specific seismograms as a function of frequency. A number of models developed using a variety of methodologies are available for testing. These include the WENA/Unified model of Eurasia (e.g. Pasyanos et al 2004), the global CUB 2.0 model (Shapiro and Ritzwoller, 2002), the partitioned waveform model for the Mediterranean (van der Lee et al., 2007) and stochastic models of the Yellow Sea Korean Peninsula region (Pasyanos et al., 2006). Secondly, we are extending our Cartesian an elastic finite difference code (WPP of Nilsson et al., 2007) to model the effects of free-surface topography. WPP models an elastic wave propagation in fully 3D earth models using mesh refinement to increase computational speed and improve memory efficiency. Thir, Proceedings of the 30th Monitoring Research Review: Ground-Based Nuclear Explosion MonitoringTechnologies, 23-25 Sep 2008, Portsmouth, VA p231-240. Sponsored by the National Nuclear Security Administration (NNSA) and the Air Force Research Laboratory (AFRL). The original document contains color images.

Earthquake source parameters underpin several aspects of nuclear explosion monitoring. Such aspects are: calibration of moment magnitudes (including coda magnitudes) and magnitude and distance amplitude corrections (MDAC); source depths; discrimination by isotropic moment tensor components; and waveform modeling for structure (including waveform tomography). This project seeks to improve methods for and broaden the applicability of estimating source parameters from broadband waveforms using the Cut-and-Paste (CAP) methodology. The CAP method uses a library of Green's functions for a one-dimensional (1D, depth-varying) seismic velocity model. The method separates the main arrivals of the regional waveform into 5 windows: Pnl (vertical and radial components), Rayleigh (vertical and radial components) and Love (transverse component). Source parameters are estimated by grid search over strike, dip, rake and depth and seismic moment or equivalently moment magnitude, MW, are adjusted to fit the amplitudes. Key to the CAP method is allowing the synthetic seismograms to shift in time relative to the data in order to account for path-propagation errors (delays) in the 1D seismic velocity model used to compute the Green's functions. The CAP method has been shown to improve estimates of source parameters, especially when delay and amplitude biases are calibrated using high signal-to-noise data from moderate earthquakes, CAP+., Presented at the Conference on Monitoring Research Review: Ground-Based Nuclear Explosion Monitoring Technologies (30th), held in Portsmouth, VA, on 23-25 Sep 2008. Published in the proceedings of the conference, p408-418, 2008. Proposal no. BAA06-04. Prepared in cooperation with Lawrence Livermore National Laboratory, and with URS Group, Inc. Sponsored in part by the National Nuclear Security Administration (NNSA). The original document contains color images.

In this project, we have exploited unique and open-source seismic datasets to improve seismic monitoring across the Middle East (including the Iranian Plateau, Zagros Mountains, Arabian Peninsula, Turkish Plateau, Gulf of Aqaba, Dead Sea Rift) and the Horn of Africa (including the northern part of the East African Rift, Afar Depression, southern Red Sea, and Gulf of Aden). The data sets have been used to perform two related tasks. (1) We have determined moment tensors, moment magnitudes, and source depths for regional events in the magnitude 4.0 to 6.0 range. (2) These events have been used to characterize high-frequency regional phase attenuation and detection thresholds, especially from events in Iran recorded at stations across the Arabian Peninsula. In the first part of this project, seismograms from earthquakes in the Zagros Mountains recorded at regional distances have been inverted for moment tensors, and source depths for the earthquakes have been determined via matching regional waveforms using a grid search algorithm and forward modeling of teleseismic depth phases. Early studies of the distribution of seismicity in the Zagros region found evidence for earthquakes in the upper mantle. But subsequent relocations of teleseismic earthquakes suggest that source depths are generally much shallower, lying mainly within the upper crust. For all of the events that have been studied, source depths lie within the upper crust. And the events all have thrust mechanisms with E-W or NW-SE striking nodal planes. In the second part of this project, the source mechanisms for these events have been used to characterize high-frequency (0.5-16 Hz) regional phase attenuation and detection thresholds for broadband seismic stations in the Arabian Peninsula, including International Monitoring System (IMS) stations and stations belonging to the Saudi Arabian National Digital Seismic Network., Presented at the Conference on Monitoring Research Review: Ground-Based Nuclear Explosion Monitoring Technologies (30th), held in Portsmouth, VA, on 23-25 Sep 2008. Published in the proceedings of the conference, p190-200, 2008. Prepared in collaboration with Lawrence Livermore National Laboratory. Sponsored in part by the National Nuclear Security Administration (NNSA). Proposal no. BAA05-07. The original document contains color images.

In an effort toward improving current seismic-velocity models for the Middle East, the initial Step of building a high-quality database of recordings from well characterized sources is essential. This high-quality database of recordings, or broadband waveforms, encompassing the region from the Mediterranean to the eastern edge of Tibet, will be the primary measure for evaluation and improvement in the iterative, waveform-based, adjoint-inversion process. Adjoint inversions, as with any waveform-based method, require precise estimates of the location, especially depth, and the faulting parameters or moment-tensor elements of the seismic sources. Imprecise locations and/or source terms will result in real changes in the amplitude and the phase of the incoming wave train; when close to a nodal plane, these changes can occur very rapidly with incoming azimuth. These errors will directly map into the resulting velocity model as uncertainty or biased values. The initial database included about 250 well recorded, regional events in the Middle East from 1990 to 2007 with magnitudes Mw> 5.5. To reduce errors in the source-depth and faulting parameters, the events were evaluated using two complementary sets of data, a regional data set contained within the Middle East and a global teleseismic data set. Regional data were first utilized at long period, omega > 20s, using the waveform-based methods of Pasyanos et al. (1996) to determine the moment tensor elements. Then shorter periods, omega > 5 s, were added, following Zhao and Helmberger (1994), to further assess errors in source depth and propagation effects by splitting apart the longer period surface waves from the shorter period, depth-sensitive Pnl waves. Problematic, or high-error, stations and paths were higher analyzed to identify systematic errors with unknown sensor responses and complex wave propagation regions., Presented at the Conference on Ground-Based Nuclear Explosion Monitoring Technologies (30th), held in Portsmouth, VA, on 23-25 Sep 2008. Published in the proceedings of the conference in Monitoring Research Review, v1 p261-267, 2008. Prepared in cooperation with Lawrence Livermore National Laboratory, and Princeton University. The original document contains color images. All DTIC reproductions will be in black and white.

Comprehensive nuclear-test-ban monitoring in terms of location and discrimination has progressed significantly in recent years. However, the characterization of sources and the estimation of low yields remains a particular challenge. As the recent Korean shot demonstrated, we can probably expect to have a small set of teleseismic, far-regional and high-frequency regional data to analyze in estimating the yield of an event. Since stacking helps to bring signals out of the noise, it becomes useful to conduct comparable analyses on neighboring events, earthquakes in this case. If these auxiliary events have accurate moments and source descriptions, we have a means of directly comparing effective source strengths. Although we will rely on modeling codes, 1D, 2D, and 3D, we will also apply a broadband calibration procedure to use longer periods (P >5 s) of waveform data to calibrate short-period (P between 0.5 to 2 Hz) and high-frequency (P between 2 to 10 Hz) as path-specific station corrections from well-known regional sources. We have expanded our basic cut-and-paste (CAP) methodology to include not only timing shifts but also amplitude (f) corrections at recording sites. The name of this method was derived from source inversions that allow timing shifts between "waveform segments" (or cutting the seismogram up and re-assembling) to correct for crustal variation. For convenience, we will refer to these f-dependent refinements as CAP+ for (short period, SP) and CAP++ for still higher frequency. These methods allow the retrieval of source parameters using only P-waveforms where radiation patterns are obvious as demonstrated in this report and are well suited for explosion P-wave data. The method is easily extended to all distances because it uses Green's function although there may be some changes required in t* to adjust for offsets between local vs. teleseismic distances., Presented at the Monitoring Research Review (29th): Ground-Based Nuclear Explosion Monitoring Technologies held in Denver, Colorado on 25-27 September 2007. Published in the Proceedings of the Monitoring Research Review (29th): Ground-Based Nuclear Explosion Monitoring Technologies, p80-90, 2007. Sponsored by the National Nuclear Security Administration (NNSA) and the Air Force Research Laboratory (AFRL). Performed in cooperation with the California Institute of Technology, Padadena, CA.The original document contains color images.

In this project, we are exploiting several seismic data sets to improve U.S. operational capabilities to monitor for low yield nuclear tests across the Middle East (including the Iranian Plateau, Zagros Mountains, Arabian Peninsula, Turkish Plateau, Gulf of Aqaba, Dead Sea Rift) and the Horn of Africa (including the northern part of the East African Rift, Afar Depression, southern Red Sea and Gulf of Aden). The data sets are being used to perform three related tasks. (1) We are determining moment tensors, moment magnitudes and source depths for regional events in the magnitude 3.0 to 6.0 range. (2) These events are being used to characterize high-frequency (0.5-16 Hz) regional phase attenuation and detection thresholds, especially from events in Iran recorded at stations across the Arabian Peninsula. (3) We are collecting location ground truth at GT5 (local) and GT20 (regional) levels for seismic events with M is greater than 2.5, including source geometry information and source depths. Towards meeting these objectives, seismograms from earthquakes in the Zagros Mountains recorded at regional distances have been inverted for moment tensors, which have then been used to create synthetic seismograms to determine the source depths of the earthquakes via waveform matching. The source depths have been confirmed by modeling teleseismic depth phases recorded on Global Seismographic Network (GSN) and International Monitoring System (IMS) stations. Early studies of the distribution of seismicity in the Zagros region found evidence for earthquakes in the upper mantle. But subsequent relocations of teleseismic earthquakes suggest that source depths are generally much shallower, lying mainly within the upper crust. All of the regional events studied so far nucleated within the upper crust, and most of the events have thrust mechanisms., Presented at the Monitoring Research Review: Ground-Based Nuclear Explosion Monitoring Technologies (29th) held in Denver, CO on 25-27 Sep 2007. Published in Proceedings of the Monitoring Research Review: Ground-Based Nuclear Explosion Monitoring Technologies (29th), p424-433, Sep 2007. The original document contains color images.

The use of 3-D geophysical models to predict nuclear test ban monitoring observables (phase travel times, amplitudes, dispersion, etc.) is widely anticipated to provide improvements in the basic seismic monitoring functions of detection, association, location, discrimination and yield estimation. A number of questions arise when contemplating a transition from 1-D, 2-D and 2.5-D models to constructing and using 3-D models, among them: (1) Can a 3-D geophysical model or a collection of 3-D models provide measurably improved predictions of seismic monitoring observables over existing 1-D models, or 2-D and 2 1/2-D models currently under development? (2) Is a single model that can predict all observables achievable, or must separate models be devised for each observable? How should joint inversion of disparate observable data be performed, if required?; (3) What are the options for model representation? Are multi-resolution models essential? How does representation affect the accuracy and speed of observable predictions?; (4) How should model uncertainty be estimated, represented, and how should it be used? Are stochastic models desirable?; (5) What data types should be used to construct the models? What quality control regime should be established?; (6) How will 3-D models be used in operations? Will significant improvements in the basic monitoring functions result from the use of 3-D models? Will the calculation of observables through 3-D models be fast enough for real-time use or must a strategy of pre-computation be employed?; (7) What are the theoretical limits to 3-D model development (resolution, uncertainty) and performance in predicting monitoring observables? How closely can those limits be approached with projected data availability, station distribution and inverse methods?; (8) What priorities should be placed on the acquisition of event ground truth information, deployment of new stations, development of new inverse techniques., Presented at the Monitoring Research Review (29th): Ground-Based Nuclear Explosion Monitoring Technologies held in Denver, Colorado on 25-27 September 2007. Published in the Proceedings of the Monitoring Research Review (29th): Ground-Based Nuclear Explosion Monitoring Technologies, p70-79, 2007. Sponsored by the National Nuclear Security Administration (NNSA) and the Air Force Research Laboratory (AFRL).The original document contains color images.

We have used data from the Kuwait National Seismic Network (KNSN) to estimate the seismic structure of Kuwait using a limited amount of seismic data. First, we made surface wave dispersion measurements and calculated receiver functions from the relatively small amount of data available from the broad-band station, KBD. Models were derived from the joint inversion of teleseismic receiver functions and Rayleigh and Love fundamental mode surface wave group velocity dispersion. While both surface waves and receiver functions by themselves can be used to estimate lithospheric structure, we have successfully combined the two to reduce non-uniqueness in estimates based on the individual data sets. The resulting KUW1 model features a thick (8 km) sedimentary cover and crustal thickness of 45 km. Crustal velocities below the sedimentary cover are consistent with global averages for stable platforms. We infer upper-mantle velocities (7.84 km s-1 P-wave velocity; 4.40 km s-1 S-wave velocity) that are slightly lower than expected for a stable platform. In comparison with other crustal structure estimates for the Arabian platform to the west, the crust is thicker and the mantle is slower in Kuwait. This is consistent with the overall tectonic trends of the region that find increasing crustal thickness between the divergent plate boundary at the Red Sea and the convergent plate boundary at the Zagros Mts, as well as slow mantle velocities beneath this nearby orogenic zone. The resulting model fits the traveltimes of regional phases (Pn, Pg, Sn and Lg). Independent inversion of local earthquake traveltimes recorded by KNSN (allowing for event hypocentre relocation) results in a remarkably similar velocity structure, providing confidence that the joint inversion of receiver functions and surface wave group velocities can impose accurate constraints on crustal structure for local event location and network operations. Relocation of events in Kuwait improves the clustering of events