Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
DISTRIBUTION OF ISMS DATA WITHIN THE UNITED STATES In this chapter, we address the second charge to the panel, which involves access to ISMS seismological data by the U.S scientific community. The CTBT monitoring system will collect several other types of data (e.g. infrasound, hydroacoustic, chemical) that are not considered here. We consider the mechanisms and infrastructure required for providing broad access to the ISMS seismic data for multiple-use applications, including nuclear test monitoring, earthquake monitoring, and research efforts that support these monitoring functions. The design of the CTBT monitoring system has significant implications for the future of nuclear monitoring in the United States and for the structure of seismic monitoring and research on a wide variety of topics of vital interest to the United States. Substantial money is about to be spent on the CTBT monitoring infrastructure, and it is desirable that it be spent wisely and effectively to ensure cost-effective usage of the data for a range of applications. To enable multiple uses of the seismic data, it is important to establish convenient pathways for data access that do not interfere with the primary operations of the nuclear test-ban monitoring effort. This report proposes cost-effective strategies that will provide these pathways. The panelts approach designs the U.S. nuclear monitoring effort to take advantage of existing data archival and distribution capabilities that service seismological applications benefiting the nation. The panel strongly recommends that the U.S. ISMS-NDC coordinate its efforts with the earthquake monitoring operations of the USGS, the data distribution capabilities of IRIS, and the research and development efforts related to treaty monitoring. This concept of an ISMS-NDC with a multi-element data distribution process, rather than an isolated center servicing all fimctions, can achieve significant cost reductions and will ensure that full access to the TSMS data is sustained. 33
34 Comprehensive Test Ban Monitoring System Introduction and Background The pane} was asked to consider the following charge: Data Access. The GSE has specified that all authorized users (most likely the 1[SMS National Data Center in each participating country) have prompt electronic access (perhaps through the ISMS International Data Center) to all raw and processed data. What kind of access would best satisfy the requirements of other operational groups (e.g., for earthquake hazards and tsunami warning)? How should the data be organized (e.g., by region, station, time penod; continuous vs. event segments) and made available (e.g., access time scales minutes or days; and media- electronic or optical)? The text of this charge was circulated to the panel's liaison representatives and to numerous members of the seismological research corrununity, with a request for feedback on data access issues. The responses underscored the central importance of data access to all members ofthe earthquake monitoring and research communities and emphasized the need to incorporate planning for broad data access to near real-time and archived data in the design of the ISMS and the U.S. ISMS-NDC. Strong sentiments were expressed that the seismic data ofthe TSMS, all of which are unclassified, should be available to both the broacler research and the earthquake monitoring communities in a timely manner. The seismic data used for past and present nuclear monitoring purposes, many of which are now unclassified, are not accessible for scientific research and thus fail to achieve their maximum impact in both the nuclear monitoring and earthquake monitoring communities. The pane! believes that the new context of unclassified data collection for the ISMS provides an opportunity to implement greater usage ofthe data streams acquired for monitoring nuclear test treaties than has been the case in the past. The recommendations made below optimize multiple use of the ISMS data while furthering the primary mission of monitoring a CTBT. Agencies with an Interest in Seismic Data Because many ofthe recommendations in this report deal with the handling and ultimate use of data, it is relevant to review the various agencies with intersecting missions and uses of seismic data. Within the federal government there are several agencies involved with seismic monitoring of earthquakes and/or nuclear testing, and several agencies involved in seismic operations and research. The DOD has traditionally held major roles in research arid development in support of nuclear monitoring, mainly organized under the Air Force Office of Scientific Research (AFOSR), the Air Force
DISTRIBUTION OF ISMS DATA WITHIN THE UNITED STATES 35 Phillips Laboratory (AFPL), and the Advanced Research Projects Agency (ARPA). Actual monitoring and national verification operations, along with advanced development research, have been primarily conducted through the Air Force Technical Applications Center (AFTAC). The Department of Energy (DOE) has a long history of source-mechan~sm, regional propagation, and seismic discrimination research using data from nuclear explosions at the Nevada Test Site and western U.S. earthquakes. Beginning in late 1994, the primary mission for research and development in support of nuclear monitoring was transferred to DOE, with the ARPA seismic monitoring research and development effort scheduled to phase out over the next two years. (AFOSR and AFPL research programs wall continue and will, in combination with DOE, provide necessary scientific and technological support to AFTAC, ARPA will continue to support development of the ISMS-IDC). This structure is considered in greater detail in Chapter 5. Global and national earthquake monitoring and basic research have been supported by the USGS. It has also coordinated on data acquisition with AFTAC. Other federal organizations involved in earthquake monitoring include the National Oceanographic and Atmospheric Administration (NOAA), the Federal Emergency Management Agency (FEMA), the National Institute for Standards and Technology (NIST), and the Nuclear Regulatory Commission (USNRC). The National Science Foundation (NSF) also supports basic seismological research on earthquakes and earth structure. It funds the Incorporated Research Institutions for Seismology (IRIS), which collects, archives, and distributes seismic data from a global array of permanent stations, temporary regional networks, and portable instrumentation. IRIS has also received funding through the DOD explicitly for data acquisition and research related to nuclear monitoring efforts. The recommendations that are made in the next section address both the seismological waveform information collected by the TSMS and the parametric measurements, such as arrival tunes and associated event bulletins. We briefly consider the nature of these forms of seismological information to provide a context for the recommendations. Seismic Waveform Data All sources of rapid change of strain energy in Earth produce seismic waves that propagate throughout the planet. A high-quality recording of ground motion can capture information about natural phenomena such as earthquake faulting, tidal motions, volcanic eruptions, and large landslides, as well as capturing human-induced vibrations, such as those from nuclear arid chemical explosions. Thus, seismograms (recordings at
36 Comprehensive Test Ban Monitoring System fixed locations of ground vibrations as a Unction of time) provide the basic information required for nuclear test aIld earthquake monitoring, disaster response to earthquakes, tsunamis, and volcanic eruptions, natural resource development, and basic research into Earth's structure and tectonics processes. Seismic waveforms are generally quite complex, and the full information content of the signal cannot be reduced to simple parametric measurements, such as the arrival times and amplitudes of discrete phases. The overall wave shape contains valuable information about the source process that generated the disturbance and the interaction of the radiated wavefield with Earth's structure. By designing ground motion sensors that record a wide range of frequencies (broadband sensors), seismologists increase information content in the recorded signals. For about a century, seismologists have been developing source and wave propagation theory and analytic methods to extract information from the completed seismogram. This is now an advanced quantitative science, and, with the latest generation of seismic instrumentation providing nearly complete recording of all ground motions at a given site, every broadband seismogram can provide extensive information about source, path, and receiver- site effects. For example, seismic waveforms can be used to locate events (using the characteristic sequence and amplitudes of arrivals recorded at varying distances from a source), to determine the orientation of an earthquake fault and the sense of shearing motion during the rupture, to image the variable slip on the fault surface, and to quantify the total energy release during the event. Modeling broadband seismograms identifies characteristics of earth structure, like the heterogeneity of the crust, that influence the signals from a surreptitious nuclear test. For seismograms recorded at regional distances of from ~ 00-1,200 km, which involve complex reverberations in the shallow crust, the relative amplitudes of different arrivals as a function of frequency are among the strongest diagnostics of the source type (e.g., OTA, ~ 9881. These regional signals are of great importance for monitoring a CTBT, because for the smallest events of concern they may be the only data available. Automation is essential for U.S. CTBT monitoring, as there will be thousands of events that must be detected, located, and identified each year. The advanced state of waveform analysis and processing is such that the longest delay in determining the faulting orientation of large earthquakes around the world tends to be the propagation delay, that is, the time it takes for the seismic waves to arrive at a sufficient number of stations to perform a stable analysis. Rapid access to seismograms from many stations makes it possible to automate many aspects of routine nuclear test and earthquake monitoring. Most research applications do not require real-time access to waveform data, but they do require the ability to retrieve diverse segments of past recordings. The waveforms are needed because no simple standardized set of archival parameters will
DISTRIBUTION OF ISMS DATA WITHIN THE UNITED STATES 37 service all applications. In some cases, only a few seconds of a given recording may be analyzed, while in other cases many hours of the same record may be used in studying either the source or the earth structure. This has prompted the development of the fRIS-DMS archive of continuous broadband seismic data, which has the ability to service individually tailored data requests. With the role that frequency dependence of waveform energy and path effects play in discriminating nuclear explosion signals from quarry blasts and earthquakes, availability of archives of waveforms from previous regional events is also as critical for nuclear monitoring. This is particularly true when monitoring a region with no prior history of large explosions. Thus, archives of seismic waveforms play a critical role in the nuclear monitoring arena as well. Broadband three-component waveforms form the primary data base now used in basic earthquake source and earth structure investigations. Typically, broadband data are recorded at relatively isolated observatory stations, although arrays of portable and semi- permanent broadband stations have been deployed recently. The more extensive the global coverage provided by broadband recordings, the more detailed the information about sources and deep structure that can be retrieved. For some applications, only a modest number of global stations are needed, possibly providing data in near real-time (this would apply, for example, to tsunami warning systems)' but almost all seismic analysis procedures are enhanced by increasing the number of observations, as long as the quality of those observations is high. Much of the enthusiasm in the research community for access to the ISMS broadband data stems from the enhanced research potential provided by the increased numbers of broadband recordings that will be available from stations located around the world for each earthquake or explosion event. The ISMS also includes a number of sites with arrays of closely spaced, high-frequency instruments in addition to a centrally located, three-component broadband instrument. By combining the data from the high-frequency instruments in various ways, analysts are able to detect signals from smaller events, identify the type of seismic wave producing the signal, locate the events more accurately, and in some cases, associate overlapping signals from multiple sources with the proper source. The instrumentation at marry of these stations is particularly appropriate for the analysis of the higher-frequency wave field generated by small events. Conversely, small arrays are of limited incremental use for studies of earthquake sources and the structure of Earth which require long-period waves. For effective monitoring of a CTBT, knowledge of crustal structure and earthquake characteristics on a global basis is very important. For example, research on the crustal structure under ISMS stations requires that the research community have access to the data from the ISMS stations. Some ofthe many ways in which providing
38 Comprehensive Test Ban Monitoring System broad access to the ISMS data will enable earthquake monitoring and basic research activities that would improve CTBT monitoring are by: · allowing research on event detection, discrimination, and yield estimation to be conducted using the same data as employed in the operational environment; · improving our knowledge of earth structure and hence improving the accuracy of event locations, and aiding in regional characterization and in the resolution of ambiguous events; · augmenting the data available for basic research on regional crustal and upper mantle structure in regions of importance for CTBT monitoring; and · augmenting the data available for research on earthquake mechanisms, source depths, and scaling properties in regions of-importance for CTBT monitoring. All these activities require the use of waveform data. In addition, the parameter data produced by the operations of the ISMS will be important for preparation of event bulletins, which is discussed next. Seismological Event Bulletins Both nuclear monitoring and earthquake monitoring involve the preparation of bulletins of events. These bulletins are lists of seismic events, usually arranged chronologically, that give at a minimum the origin time, the event location (latitude, longitude, depth), and one or more seismic magnitudes, all determined from an analysis of the seismic wave arrivals observed at and reported from stations around the world. Some bulletins also include, for each listed event, the arrival times of seismic waves detected at each station. Traditionally, many seismograph stations have reported arrival times regularly to the organizations that publish bulletins. Various types of global bulletins have been developed over the years. The International Seismological Centre (ISC), located in Newbury, England, publishes the most complete and most accurate bulletin of global seismicity. The ISC bulletin appears about two years in arrears and is based upon infonnation (such as the observed arrival times of various seismic waves) contributed currently from more than i,800 seismographic stations around the world (see Figure 2.5), most of them recording in analog formats such as ink on paper or photographs. At present, the ISC bulletin is not complete even at magnitude 5.5 in some remote parts of the southern hemisphere. (paper given by Robin Adams, IRIS workshop, 19911. The USGS National Earthquake Information Center (NETC) publishes global bulletins known as the Quick Epicenter Determination (QED) and the Preliminary
DISTRIBUTION OF ISMS DATA WITHIN THE UNITED STATES 39 Determination of Epicenters (PDE). The QED is first produced about an hour in arrears for a very limited number of earthquakes, and more generally about a week in arrears. The PDE appears about a month in arrears. These bulletins are produced on different time scales in order to service different needs of the earthquake monitoring community. The time lag allowed for any bulletin preparation determines the number of data that can be used to detect and locate events, given that many stations are in remote areas with limited communications. For ~ 99 I, the NElC located ~ 6,5 ~ 6 events, including i,585 events of seismic magnitude 5 and above, and 4,372 events with magnitudes from ~ 4.9. That year, the TSC located i,373 events previously unidentified in bulletins utilizing fewer stations, 34 with magnitudes ~ 5, 209 with magnitudes 4.4 4.9, 28O with magnitudes 4.0~.4, and 850 with magnitudes ~ 4. The primary stations envisioned for GSETT-3 have been estimated (CD/1254) to have a threshold detection capability in the magnitude Parke below 3 for parts of Eurasia and North America, above magnitude 3.4 in some continental areas of the southern hemisphere, and above magnitude 3.8 in parts of the southern oceans. From these detections, plus additional data that may be requested from auxiliary stations as deemed necessary to improve location estimates, the ISMS-IDC will obtain automated event locations that will be reviewed and, if necessary, corrected by an analyst. The GSETT-3 IDC is now publishing a bulletin of global seismicity, known as the Reviewed Event Bulletin (REB), two days in arrears. This production schedule is shorter than those of the TSC and the USGS (except for a limited number of earthquakes studied promptly by the USGS, together with TRIS). For treaty-monitoring purposes, it will be important to be able to examine the region of a suspicious seismic event as soon as possible, using available methods, and if suspicions persist, to request and carry out an · . - on-s~te inspection. The GSETT-3 REB differs from current USGS and ISC bulletins in ways other than timeliness of production. The REB is based upon the ISMS-IDC's own analysis of digital seismograms communicated in near real time, rather than the ISC's practice of analyzing measurements of wave arrival times made at the contributing stations. The REB uses fewer stations, although some are arrays, to locate events than either the USGS or the TSC. In the early days of GSETT-3, the REB may give event locations that in many cases wall not be as accurate as the later-published USGS and ISC bulletins, but as the GSETI network is calibrated, the quality watt improve. The REB may have more uniform global coverage than the USGS and ISC, but the latter's bulletins are likely to have improved coverage in certain areas. Certainly it may be expected that in parts of the world where strong national programs exist to study earthquake hazard (e.g., China' Japan, Mexico, and the United States), these latter programs will provide seismicity bulletins far superior to the REB in all attributes except timeliness. As a result, the REB includes events that the USGS and ISC now miss; and conversely, the USGS and the
40 Comprehensive Test Ban Monitoring System ISC bulletins include events that are not in the REB. Comparison of the various bulletins should enhance all. ]:SMS-NDCs may also contribute supplemental data to the ISMS-IDC. Such data could include seismicity bulletins based upon various regional, national, and international station networks, and the seismic-wave arrival times from analog and digital stations upon which these bulletins are based. For the GSETT-3 experiment, only seismicity bulletins will be provided, and these will be used in assessing the quality and completeness of the REB. However, it does not appear that the ISMS-IDC plans to use these supplemental data in its bulletin preparation because they will usually arrive too late to assist in production of the REB. Furthermore, there does not appear to be a systematic plan to compare the bulletins. Given the current timing constraints, supplemental data and ISC/USGS operations cannot assist GSETT-3 in production of the REB. However, the data gathered by GSETT-3 could greatly assist the ISC and the USGS in ways that could have a positive impact on the general effort to improve CTBT monitoring. If the REB and its underlying wave picks were made available to the ISC and the USGS, this information could be combined with other data available to those organizations and they could provide more accurate locations than the REB for all seismic events above about magnitude 4.5 (about 2,300 events per year; see Table 4.~) and perhaps almost all events down to magnitude 4 (about 7,100 events per year). As a long-term goal, complete coverage on continents down to magnitude 3 may be achievable, using current and planned stations. A comprehensive global seismicity bulletin that emphasizes completeness and accuracy of location rather than speed of publication would improve CTBT monitoring by: · providing location estimates generally superior to those of the REB, thus supporting evaluation of the REB and the preparation of guidance on how to improve it; · locating events not included in the REB, thus allowing an evaluation of the REB threshold in different regions; · providing an archive of accurate event locations useful for prompt interpretation of new seismic locations in the REB as they accumulate; · supporting special studies of seismicity in different regions of Earth that may be of CTBT concern; · improving our knowledge of earth structure and hence improving the accuracy of event location by the REB; and . aiding in regional characterization and in the resolution of ambiguous events.
DISTRIBUTION OF ISMS DATA WITHIN THE UNITED STATES TABLE 4.1. Seismic magnitude Number of earthquakes per year (and per day), worldwide, above each magnitude* 4.5 4.0 3.5 3.0 2.5 2,300 (6) (19) (52) (1 86) 7,100 19,000 68,000 209,000 (572) * Based on Ringdal (1985). 41 At present, most earthquake bulletins intentionally exclude quarry blast information to avoid contamination of the natural seismicity information. For CTBT monitoring purposes, information about quarry blasts is of great importance, and the REB will include many such sources. It will be a major task ofthe national verification activity to identify quarry blasts and to ensure that none of the explosions are nuclear tests. The earthquake monitoring community could make a significant contribution to CTBT monitoring by determining quarry blast locations on regional and global scales, perhaps producing a separate bulletin for such events. This would require modification of existing procedures in which many station operators screen out quarry blast information, and there would be a nontrivial cost for the additional operations; however it would provide additional information for identifying the many quarry blasts detected by the ISMS, especially those on U.S. territory. There are additional general reasons to develop an improved global bulletin using all the data available to the TSC and the USGS-NElC, augmented by data from the GSETT-3 and the subsequent TSMS. A global bulletin is the primary database summarizing seismic activity for many interested users outside seismology in geophysical research and in quantitative estimation of seismic hazard. (Far more scientists and engineers use seismicity bulletins than use seismograms directly.) For the seismological research community, an improved bulletin with the goal of complete coverage above a certain (Iow) magnitude threshold would be important for research on earthquake prediction and hazards. It would also focus attention on the need for quiet sites, highly reliable stations, and appropriate station siting all issues of concern for explosion monitoring.
42 Comprehensive Test Ban Monitoring System GSETT-3 is essentially focused around the effort to produce a global bulletin of seismicity only two days in arrears. If the data generated by GSETT-3 for this purpose (in particular the wave picks) were made available to organizations now publishing bulletins weeks and years in arrears, the outcome would likely be significantly improved accuracy of event location and improver! global coverage down to lower magnitudes. Since seismicity bulletins are among the most basic and important databases in geophysics and in the study of natural hazard reduction, many scientists and engineers and their clients, including the general public as well as the CTBT monitoring community, would benefit. Fundamental Guidelines for Data Access Issues Effective seismic monitoring of a CTBT requires the detection, location, and identification of underground nuclear explosions with high confidence and a low false- alarm rate. in this context, once an earthquake is clearly identified as such there is no further immediate interest in its signals. (In the longer term, an archive of such signals can assist future discrimination efforts by providing comparison recordings from the same region.) However, these same recordings are of great value for other applications, such as earthquake monitoring, analysis of the earthquake faulting process, and analysis Of structure of Earth. If the instrumentation operated by the ISMS includes appropriate dynamic range and bandwidth, the recorded signals are certain to be useful for many purposes in addition to routine screening to detect nuclear explosion signals. (This was the basic premise underlying the recornrnendations in Chapter 3.) The ISMS data quality, distribution of stations, digital format, broadband response, large dynamic range, and timely electronic access all contribute to the potential value of the data. In order for the data to Fife this potential for both the monitoring and the broader seismological communities, they must reside in readily accessible archives. The pane] feels strongly that it is in the interest of efficient use of resources to ensure that the ISMS data be accessible to U.S. scientists by means that maximize their usefulness to all efforts of national interest, to the extent possible without compromising the basic mission of the TSMS. The combined user community should attempt to integrate all international seismic data acquisition, archiving, distribution, and bulletin preparation efforts in a way that benefits all of the potential users of seismic data and strengthens monitoring capabilities in both the short and Tong term. The pane} concludes that efficient integration of the ISMS and ISMS-NDC with existing facilities for earthquake monitoring and distribution of data to the seismic research community can provide benefits to the nuclear test and earthquake monitoring communities and to the research efforts that support them.
DISTRIBUTION OF ISMS DATA WITHIN THE I]NITED STATES 43 As the ISMS is developed, it is critical to recognize that there are many existing international arrangements for open access of seismic data, and these should not be undermined by the political sensitivities associated with nuclear monitoring efforts. The pane' recommends that steps be taken to ensure that the development of the ISMS does no! result in a reduction of existing capabiiifies of the U.S. scientist c community The United States should support the policy that incorporation of existing stations into either primary or auxiliary station affiliation with the ISMS should not result in restriction of the current access to these or other stations. This requires that there be either open access to the complete data streams through the ISMS or that arrangements be made for access via previous procedures. Furthermore, the United States should support a policy that establishes similar procedures for new stations. In addition, a nation's participation in the TSMS should not introduce barriers to obtaining ciata from other stations previously operating in that country. This issue should be addressed in the preparation of the protocol to the treaty. For the primary stations, access through the ISMS would be advantageous because the rapid, centralized collection of the data will provide signals from a worId-w~de network of high-quality stations in near real time from a single source. This could benefit disaster mitigation and response efforts as well as facilitate seismic research into a number of areas of real-time seismic data processing, with attendant monitoring, scientific, and societal benefits. For the auxiliary stations, access to the segmented data from the ISMS will also increase rapid access to many stations, but it is important that alternate means of retrieving the complete continuous data from these stations be maintained or, in the case of new stations, established. This will require coordination between operators of the auxiliary stations and the nuclear monitoring community, which will be investing in these stations, to ensure that on-demand access is available and that the data satisfy the operational requirements for the auxiliary network. Historically, many ofthe seismic data collected by the United States for nuclear monitoring have not been accessible to the scientific community, even though the data were unclassified. The ISMS stations and their data will be unclassified even though they will be used for monitoring purposes. It is important that the ISMS data be available to the broader seismic community because this will benefit many activities, including CTBT monitoring. Development of new methods of analysis and testing of nuclear monitoring procedures will be facilitated by access to the actual data that are used in the monitoring operation. This access will enhance the interactions between the nuclear monitoring research and operational communities. This fact has been amply demonstrated in other contexts in the past.
44 Comprehensive Test Ban Monitoring System The research community can also play a part in the confidence-building process that is an essential element in the justification of the ISMS. These researchers will be advisors to their governments and will provide important independent checks and balances on the operations of the monitoring system, as well as sources of insight into the geophysical properties of regions of Earth, the nature of specific events of interest, and monitoring methods in general. In addition, the broader the user community is, the better the feedback about quality-control issues and instrumentation problems. Such problems are often revealed in the course of analysis of recordings for large earthquakes, which may be ignored in the national verification effort. The pane! recommends that the ·SMS and nuclear monitoring communities adopt a clear commitment to provide ready access to all seismic data collected by the ·SMS, and that language to this effect be inserted into the Protocol to the Tread The ISMS-IDC will provide to the I:SMS-NDCs data that will directly service national nuclear monitoring applications and other activities. The pane' recommends that all the ·SMS data received by the US ·SMS-NDC be made available to the earthquake monitoring agencies and the scientist c community in a timely manner, as well as to the nuclear monitoring operation To accomplish this, the U.S. TSMS-NDC must be committed to interfacing with both the broader scientific community and the nuclear monitoring community. Language to this effect should be placed in the tasking requirements of the U.S. NDC operating organization or organizations. The pane! recommends that the U.S Government establish a multiagency advisory committee, wit/' representation from the earthquake monitoring and basic research communities fofacilitate interagency data transmission audits address cost issues The current plan for the ISMS-NDC wait maintain an archive of all broadband arid array data from the primary and auxiliary stations; however, this will probably not be a readily accessible online archive. The pane! recommends that the ISMS-NDCforward the data streams that are of greatest interest for other applications from the ·SMS-IDC to appropriate earthquake monitoring facilities. The waveform data should be accompanied by associated calibration and station parameter information Reformatting the data to achieve a single archival format is highly desirable. The pane! recommends that there be no restrictions on the availability of primary and auxiliary data (such as limiting availability to those data provided bythe
DISTRIBUTION OF ISMS DATA WITHIN THE UNITED STATES 45 U.S.-operated stations); aR ·SMS clata s/`ould be made available. Language to this effect should be included in the CTBTProtocol. Although a major interactive data distribution system could be established at the ISMS-NDC to service all user requests for the nuclear monitoring data, this is likely to be very costly and would replicate existing data distribution capabilities. In addition, there is concern that the responsibility for data distribution could be subordinated to the nuclear monitoring operations, making data access difficult in practice. Procedures for minimizing the data distribution burden of the TSMS-NDC are discussed in the next section. The pane! has attempted to identify cost-effective pathways that both ensure data access and minimize duplication of effort. It is also important to consider the facts that the ISMS system will not operate alone and that there will continue to be other seismic data collected that will prove valuable for treaty verification efforts. It is a reality today that many more seismographic stations are operational and capable of contributing significant data on particular seismic events than are included in lists of proposed primary and auxiliary stations. It has been a common practice by our own national verification group at AFTAC to seek additional data beyond the USAEDS networks in the resolution of problem events. The event of interest might have been recorded well by several stations in a regional or national network where only one or two ofthose stations are designated ISMS stations, hence critically important data may exist beyond those that the IDC wall normally access. The open access of all stations, from both the nuclear monitoring and earthquake monitoring arenas, is to be strongly encouraged. In presentations at the Conference on Disarmament in Geneva, the United States has explicitly recognized that "other seismological resources" (CD/NTB/WP.96), beyond those associated with specific national and international monitoring systems, can and should contribute to CTBT verification. There are therefore accepted operational reasons, as well as reasons stemming from support of research and development activities, for the United States to pursue practical methods of data access to all seismographic stations around the world that meet minimal criteria. The growth of digital seismographic installations in numerous countries, together with expected reductions in the cost of communications in fixture years, could mean that monitoring of all types of seismic activity will greatly improve · - In many regions. The United States can help tap into these resources by supporting the develop- ment of simple communications hardware and software and by encouraging a policy of open data access among national and regional seismographic networks in different countries. Seismologists have a tradition of freely exchanging data that goes back decades. Although the IDC will rely upon the ISMS primary and auxiliary networks for
46 Comprehensive Test Ban Monitoring System routine analysis, access to additional stations on an ad hoc basis is likely to be the key to a better understanding of what would otherwise be problem events. Agency-Specific Recommendations Concerning Data Access Issues · We now consider issues associated with each type of data stream arriving at the U.S. TSMS-NDC. This is done in the context of fundamental design of the U.S. ISMS-NDC and of specific agency activities, with the aim of identifying pathways that will implement the generic recommendations of the previous section. The seismological methods for monitoring a CTBT will be developed in three different organizational contexts. These will involve an international process associated with the treaty negotiation itself, a national process, different in each country, that defines the responsibilities of operational agencies (such as AFTAC and the USGS in the United States), and a broader seismological context involving organizations doing seismology that have no formal responsibility to report to the ISMS or to the nuclear monitoring agency. The problem of explosion monitoring and the necessary research and development efforts are very different as seen from these three organizational contexts. The pane! has addressed the data access issue from the broader seismological community perspective, emphasizing the way that advances in the general seismological community will have a positive and significant impact on explosion monitoring. There are two basic strategies that could be pursued for the ISMS-NDC, involving either a stand-alone center that has all data acquisition, processing, and distribution capabilities, or a center that uses existing distnbuted capabilities to help provide the required functions. For an autonomous center to provide open access to the very large and diverse TSMS data set in its archive, the commitment of significant resources and personnel to service highly variable data requests from the earthquake monitoring and general research communities would be required. The diversity of such data usages should not be underestimated. Even with extensive computer automation and massive data storage capabilities, such a level of activity would have a significant impact on operational activities at the ISMS-NDC. For example, the IRIS-DMS has distributed more than a terabyte of individually tailored data products, spanning the multiyear interval of its on-line data. These resources and procedures could make a major contribution to the distribution of the ISMS data, thereby reducing costs for the ISMS-NDC and facilitating user access. A concern expressed by many representatives of the earthquake monitoring and general research communities in the United States is that if the treaty monitoring agency nrnvides this service. it will not attach a high priority to providing access to the current
DISTRIBUTION OF ISMS DATA WITHN THE UNITED STATES 47 and archival data. At the ISMS-NDC, it is likely that the on-line portion of the archive will be limited to the most recent year of records (as will be the case for the GSETT-3), and there may be difficulties in obtaining access to earlier data that has been placed in off-line storage. The concern is that when fiscal limitations are imposed, support for general access to the ISMS data will likely be the first area compromised. Past experience has indicated that operational organizations have to focus on narrow goals, which can poorly serve the broader seismological community (and, in the long run, the U.S. efforts in explosion monitoring). To some extent, this issue cart be resolved by explicit tasking of the U.S. TSMS-NDC to ensure that data access is sustained. Under the structure we envisage for the ISMS-NDC, much of the servicing of data requests could be provided through existing mechanisms, reducing the need to invest in parallel distribution services. In addition, because there are likely to be multiple agencies involved in the data acquisition process, the distributed approach will naturally provide the necessary coordinated effort. The classified national verification functions would be structurally separate from the data distribution process to address data security issues. The pane] believes that a self-contained ISMS-NDC would probably be more costly and provide less access to the ISMS data, than an ISMS-NDC with distributed functions requires several levels of interagency coordination. Since the prototype ISMS-NDC explicitly involves interagency coordination between AFTAC and the USGS on data collection, it should be straightforward to coordinate on data distribution arid multiple utilization of ISMS data. The continuous three-component broadband data from the primary stations are certainly of interest to earthquake monitoring groups, such as the USGS and tsunami warning systems, near real-time access to selected stations is necessary for these earthquake monitoring applications. The pane! recommends that the US ·SMS-NDC make continuous primary station broadband data available in near real time to earthquake monitoring agencies; the data should be archived at the ·RIS-DMS The continuous broadband data should be provided with no windowing. This approach duplicates the archive of the continuous broadband data (only a small fraction ofthe total ISMS data), but it will allow data to be stored in a permanent on-line database whose primary mission is to make data available to the broad community. In addition, this approach exploits the vast existing infrastructure of the TRIS-DMS for servicing extensive user requests for broadband data. The TRIS-DMS has a commendable record of promptly supplying data in response to all requests, including those Tom non-IRIS members. The ISMS-NDC watt thereby not need to directly service the user requests for broadband ISMS data, removing a major distribution burden (far
48 Comprehensive Test Ban Monitoring System greater than the expected demand for array data, at least in the short term) and reducing its data security issues. Archiving the continuous broadband data at the {RIS-DMS will assemble broadband data from many sources in a single data base and will also allow further standardization of data formats arid instrument response information, greatly facilitating user access to the continuous global broadband seismic data. It is critical that data quality-control information from the TSMS accompany the broadband data. Otherwise, redundant and expensive quality control would have to be performed. The costs for the broadband data transmission and archival should be shared by earthquake monitoring agencies and earthquake and nuclear explosion research agencies in proportion to their projected usage. The pane! recommends that a prototype of the broadband data distribution system be developed in the GSETT-3 experiment. At the time this report was written, the USGS had been seeking to obtain the continuous data from primary stations, but was hampered by limited communication links with the ISMS-NDC connection. A direct connection to the prototype ISMS-IDC exists and could be used to pass data to the USGS earthquake mon~tonng activities gratis appropriate communications are installed. The segmented three-component broadband data from the auxiliary stations are also important for earthquake monitoring functions. The ISMS will facilitate rapid access to these data. This is expected to be a relatively small data set that would incur only small incremental costs beyond that for accessing the continuous primary station broadband data. Access to the continuous data at these auxiliary stations, much of which will never be collected by the TSMS but all of which will be of value to the earthquake monitoring arid research communities, should continue to be provided to the TRIS-DMC under existing or new arrangements. The pane! recommends that the U.S ISMS-NDC make the segmented auxiliary station broadband data that it acquires available in near real time to the earthquake monitoring agencies; the data should be archived at the IRIS-DMS Since many ofthe ISMS broadband stations are currently operated by members ofthe FDSN, which has data distribution agreements with IRIS, it should be politically straightforward to incorporate both the continuous primary and segmented auxiliary data from these stations into the {RIS-DMS. For U.S. and non-U.S. TSMS stations that are not associated with IRIS or the FDSN. agreements should be obtained to allow all oftheir broadband data to be archived by IRIS, as this will help service the archival and data distribution of all high-quality broadband data. It would be desirable to have the
DISTRIBUTION OF ISMS DATA WITHIN THE UNITED STATES 49 continuous broadband! ciata from any auxiliary stations not currently part of the FDSN also provided to the TRIS-DMS. The pane' recommends that the continuous data from auxiliary stations (most of which wig not be accessed routinely by file ·SMSJ should continue to be archived and distributed through existing procedures of t/'e FDSN. Operational support of the U.S. auxiliary stations should be shared by the nuclear monitoring, earthquake monitoring, and basic research agencies. The most technically and financially difficult data access issue involves the short-period primary array data, which have immense storage requirements and pose major challenges for maintaining a filthy accessible on-line database. These continuous data are of central importance for nuclear monitoring procedures. The need for short-period array data in the earthquake monitoring and basic research communities is currently rather limited but is almost certain to increase with time. Possible applications of the data include refined earthquake location in certain areas and investigations of deep earth structure using the high-quality array data. It is probable that array data initially will be most desired by the seismological corrununity involved in nuclear test monitoring research, with their access to the data being directly beneficial to CTBT monitoring research. For these scientists, an important issue will be whether the data are archived on-line, on tape, or on other off-line media. The current plan for AFTAC operations during GSETT-3 will be to maintain up to one year of data on line and older data in permanent 8-mm tape storage. There are several approaches to making these data available to the broader community, each with different cost implications. The first option is to send all of the ISMS data to the USGS or IRIS to distribute through existing procedures. This approach does not involve any direct servicing of user requests by the ISMS-NDC and would greatly reduce the burden on the center. However, transmission of all the array data to the {RIS-DMS with on-line archival and servicing of user requests would be excessively costly. Only moderate savings can be achieved by retaining segmented event windows on line, given the need to process the continuous data in order to obtain the segments. The probable high cost and redundancy of the massive archive makes this approach unattractive. Another approach is for the ISMS-NDC to take the role of extracting specified event-w~ndowed segments from the array data and either archiving these in a convenient retrievable form and directly servicing user requests for the segmented data, or passing the greatly reduced data set on to the-IRIS-DMS for archival and distribution. Guidelines for the precise data windows could be established by the research community, with preliminary feedback to the panel indicating primary interest in data for global events with magnitudes above 4.0, although in regions of possible decoupling, events as small as 2.5 are of interest. This approach also limits servicing of
50 Comprehensive Test Ban Monitoring System user requests by the ISMS-NDC, but places restrictions on the data availability and requires that ISMS-NDC resources be invested in the production of a windowed data set that may not receive extensive use. A third approach is to utilize the internal database management system of the ISMS-NDC to provide access to the array data. Here the major concerns are to minimize impact on the nuclear monitoring function and to protect the security of classified data bases at the TSMS-NDC. Having the ISMS-NDC service multiple user requests for the array data would require extensive software development and would require seismic data users to become familiar with a new data access system (in addition to the {RIS-DMS, USGS, and university data centers). The GSE, ARPA, USGS, FDSN, and IRIS have designed and tested a prototype of an automated data access system that could provide on demand requests for array data as well as service standing subscriptions for specific parameter windows. Such a system could be linked to USGS/TSMS-NDC or TRIS/ISMS-NDC interfaces to provide full access to the array data. The pane! recommends that access to all continuous array data be made available by a system that exploits existing seismic data distribution capabilities An interface to the TSMS-NDC data management system should be installed to allow on-demand requests for array data as well as standing subscriptions for specific parameter windows. If this is implemented, requests for array data would be submitted to TRIS or to the USGS in familiar formats like those used to obtain broadband or regional array data; an Uris DMS/ISMS-NDC or USGS/ISMS-NDC interface would retrieve the specified data windows. This approach will (~) minimize the external user request workload for the ISMS-NDC, (2) maintain a centralized access point for broadband and array data for the seismic research community, and (3) retain access to the entire ISMS data stream. Restricting the number of users with access to the ISMS-NDC data base simplifies data security considerations for pathways to other data sets collected by NTM, and ensures that the system performance is not degraded by multiple-user access. The cost for setting up this interface should be borne primarily by the nuclear monitoring research agencies, as most current user data requirements involving short-period array data are motivated by nuclear monitoring issues, although other applications will grow. Efficient utilization of the array archive would be facilitated if 30-minute segments of data from array beams for events above a certain magnitude (say 4.0) were routinely provided to the research community via IRIS or the USGS. This would allow the research community to assess event signal quality and to make requests for only promising data. This could be in the form of a standing subscription request.
DISTRIBUTION OF ISMS DATA WITHIN THE UNITED STATES 51 The pane! recommends that a prologue interface for accessing the array data be installed during the GSETT-3 experiment. The GSETT-3 test can provide a test ofthe ability of the NDC data management system to service both national verification functions and the external array data user community. Based on this experience, procedures for distributing the array data may have to be revised. I he precise way in which the U.S. ISMS-NDC is defined is a political issue, and there are inevitable tensions over agency roles. The panel's recommendations are directed essentially at how the ISMS-NDC will function, not at how it is intemally funded. The data from the TSMS-~:DC should have a direct pipeline into the national verification arena, which is likely to be at AFTAC. It is probable that the nonseismic data will come into the same location, making this the hub of the U.S. monitoring efforts. The pane! believes the broadband data will be passed on to the USGS and IRIS, to be merged with other global network data. It is also possible that some users will want to go directly to the ISMS-NDC, even for broadband data. Thus from one perspec- tive, the USGS and IRIS could be viewed as users of the ISMS-NDC data, while from another perspective they could be viewed as part of the archival and distribution function of the ISMS-NDC. This should not become a political Cashpoint, as the ultimate objective is to ensure long-term general access to the data in a cost-efficient manner. There is great value in consolidating the broadband data in a single unified data base, with many years of on-line data. The U.S. ISMS-NDC wait receive extensive seismic event information from the TSMS-IDC. This information could be of substantial use for earthquake monitoring efforts. The pane! recommends that the seismic event parameter information (arrival times and /'ypocentral parametersJ and final bulletin from the ISMS be made available to the earthquake monitoring agencies rapidly via file ISMS-NDC. These parameter data should include event locations and phase information as well as other event parameters determined by the IDC. The current plans for the ISMS are for such information to be provided to the ISMS-NDC within 48 hours, arid the information should immediately be made available to the earthquake monitoring agencies. The ISC produces the definitive global seismicity catalog for use by the international seismic research community, building upon the EDR of the USGS. TI`e pane' recommends that the seismic event parameter information (arrival times and I'ypocentral parametersJ produced at the ISMS- DC be transmitted by electronic means to the ·SC, to be incorporated in their f nal bulletin preparation
52 Comprehensive Test Ban Monitoring System This recommendation should involve negligible extra cost for the ISMS-TDC operation, but would almost certainly add costs to the TSC operation. Additional seismic informa- tion collected by the TSMS-TDC involving parameter data (arrival times and event locations) from regional arrays should also be made available to the earthquake monitoring program and the ISC. It is important to note that the high-quaTity earthquake bulletins prepared by the USGS and ISC will be valuable for the nuclear monitoring operations for improving station corrections, for providing background activity levels, and for providing a basis for comparison with the IDC/NDC operations. The pane! recommends that distribution of the parameter data to earthquake monitoring and bulletin preparation agencies be initiated in the GSETT-3 experiment. The recommendations given above are designect to establish a U.S. infrastructure that meets the needs of both the nuclear test and earthquake monitoring communities in the short and long term. The ISMS data set will be very large, so it is desirable to minimize redundancy in archiving the data. It is important to recognize that much ofthe substantial infrastructure necessary for providing convenient user access to the archive already exists. The TRIS-DMS has proven an effective center for archiving ant! distributing broadband seismic data in response to tailored user requests. Rather than replicate this capability. at the ISMS-NDC, it would be more efficient to transmit all the broadband data to this facility, along with necessary instrument calibration information. This can be synchronized with transmission ofthe broadband data to the earthquake and tsunami monitoring operations. Alternatively, the TDC data could be available in parallel to many users/data centers if appropriate satellite downlink protocols were established. It should be noted that there is precedent for providing parallel access to seismic data streams in the envisioned manner. The Global Telemetered Seismic Network/Ancillary Seismic Network (GTSN/ASN) data collection and distribution to AFTAC and through the USGS and IRIS provides a functioning mode} for this type of interaction. By providing access to the continuous array data through the USGS/ISMS-NDC or an TRIS-DMS/ISMS-NDC interface, the entire ISMS data set wall be available to the community, but the ISMS-NDC burden of servicing data requests will be minimized. Nothing would preclude the ISMS-NDC from also servicing requests for broadband data. The objective here is to provide guaranteed access to the data in a form that is as convenient as possible for the researchers. Such a structure will ensure that there is no deleterious effect on the nuclear monitoring operation while enabling optimal multiple use of the data. This integrated approach, involving open access to all unclassified seismic data, offers the first opportunity to work toward a rational U.S. National Seismological System servicing national needs.
