Click for next page ( 24


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 23
ISMS DATA CHARACTERISTICS This chapter addresses the first charge to the panel, involving the data characteristics for ISMS monitoring stations as proposed by the Group of Scientific Experts (GSE). After providing introductory background, the chapter presents comments and recommendations on the technical requirements of ISMS stations and follows that with a discussion of the data streams. The ISMS data will be an important contribution to the total seismic data needed by diverse parts of the research community. Conversely, much data from current networks can provide essential input to the monitoring effort. It is important, therefore, that the data characteristics of the new ISMS stations be as compatible with the broad needs of seismology in general as is consistent with the objectives of CTBT monitoring. Furthermore, timely general access to the data streams from broadband three-component instruments at the primary and auxiliary stations of the ISMS will allow the best utilization of the data in earthquake monitoring, research, and treaty monitoring activities to benefit the country. Specific suggestions about ensuring rapid and full access will be discussed in Chapter 4. Introduction and Background The panel was asked to address the following charge: Data Characteristics. The Group of Scientific Experts (GSE) has written draft requirements for an IS\IS-standard station that specify characteristics such as sample rate, passband, dynamic range, and sensitivity. They have also proposed a primary network configuration and rough requirements for signal detection, parameter extraction, and event location. What types of data (raw and/or processed) are sought by the seismological community for use in test-ban monitoring research and in other types of basic research? 23

OCR for page 23
24 Comprehensive Test Ban Monitoring System The draft requirements for an ISMS-standard three-component station were extracted from the Progress Report to the Conference on Disarmament CD/121 ~ and were circulated to the panel with a request for broad technical feedback. Our comments in this chapter are confined to the specifics of instrumentation in the context of multiuse potential of the seismic data streams. We do not address the CD strategy of a tiered seismic network of primary and auxiliary stations, but we do consider the broad implications ofthe instrument specifications that have been proposed for ISMS standard stations. It is assumed that these technical specifications apply to the primary stations and that the auxiliary station requirements may not be exactly the same. We will note some distinctions in design requirements for broadband stations versus short-period array components, given the different opportunities for noise-suppression processing. The issue of what types of data are sought by the seismological community is complex because many components of this community have distinct data requirements, and no single network can service all functions. As a consequence, an extensive and multifaceted national and international infrastructure exists for collecting local, regional, and global seismic data with varying technical specifications (see Chapter 21. This infrastructure supports research and monitoring functions associated with earthquake hazard mitigation, earthquake engineering, fundamental earthquake investigations, local and global earth structure investigations, and earthquake and tsunami warning systems, as well as nuclear test monitoring systems. Many federal, state, university, private, and international organizations are involved. Although each seismological application has its own special data requirements, there are significant intersections in requirements, and the U.S. and international seismic communities have extensive multiagency, multiuniversity, bilateral, and multinational agreements in place to take advar~tage of the entire data acquisition effort. The panel's perspective, therefore is that it is most efficient and cost effective to optimize instrument and data compatibili- ty to the extent possible to enable multiple uses of the seismic data. The ISMS operation will provide both improved real-time access to some existing international seismic data and access to totally new seismic data, which can potentially complement the existing data used in diverse seismic research and monitoring activities. From the research community's point of view, it is critical that international planning for ISMS stations be well coordinated with the existing seismological infrastructure servicing other areas of national need. Feedback to the panel indicates that the broad seismological community agrees that ISMS primary station data can contribute to many of these areas outside the treaty monitoring arena. as long as suitable access is provided. The panel's response to the first charge addresses proposed attributes of the technical instrumentation of TSMS standard stations, indicating ways in which minor modifications to the technical requirements will optimize multiple-use applications of

OCR for page 23
ISMS DATA CHARACTERISTICS 25 the seismic data streams. Then, we address the prioritization of the data generated by the ISMS system for both research and earthquake monitoring applications. Discussion of TSMS Station Technical Requirements TSMS instrument specifications emphasize the short-period end of the seismological spectrum, essential for recording small events, whereas the research community emphasizes recording a very broad spectrum of seismic signals with a dynamic range sufficient to resolve ground noise and to record, on scale, the largest signals. The pane} has not addressed the seismological objectives that drive the technical requirements for the ISMS stations, but they generally appear to be consistent with the needs of a treaty monitoring system. Fortunately, modern seismic instrumentation has expanded the bandwidth and dynamic range of commonly available sensors. ISMS technical specifications for the three-component broadband sensors to be deployed at each primary station are fairly close to those of the instrumentation deployed by the U.S. Geological Survey (USGS), Incorporated Research Institutions for Seismology (IRIS), and other members of the Federation of Digital Seismographic Networks (FDSN) for a broad range of seismological applications. The pane! strongly endorses the planned inclusion of at least one three-component broadband set of instruments in a low-noise environment at each primary station. The resulting continuous three-component broadband data from the TSMS would have maximum impact in basic research investigations. The broadband channels, with the specified passband, can readily be incorporated into the USGS's earthquake monitoring and analysis procedures, extending the data available for use in near real-time. However, the "Station Requirements for an ISMS Standard Station," listed in Table ~ of CD/121 I, are somewhat restrictive and limit broader applications of the data stream. Below, we discuss the relationship of the technical requirements listed in CD/121 ~ and indicate their relationship to the types of seismic data sought by the seismological corrunuruty. We note that the GSE has actively been considering some of the recommendations made below (based on the panel's first preliminary report), and new station requirements are proposed in more recent CD working documents, some of which accommodate our suggested charges. a. Passband 60. 02 to 20 Hid. The low-frequency cutoff of this passband wall preclude recording of very-Iong-period surface waves and free oscillations, which are of extensive use in basic research on earthquake sources and earth structure. It is now technologically straightforward to extend the low-frequency response to 0.003 Hz (the lowest frequency of seismological interest) without significant impact on the cost of the

OCR for page 23
26 Comprehensive Test Ban Monitoring System broadband instrumentation. This change would replicate the Tow-frequency response capabilities of many ofthe IRIS, USGS, and FDSN stations. Since many ofthe stations initially designated as primary or auxiliary stations for the GSETT-3 already have very broadband sensors (STS-! or KS54000) that intrinsically achieve a low-frequency response, down to 0.003 Hz, it is straightforward to broaden the specified bandwidth, but it is not possible at present to meet the full desired bandwidth using a single instru- ment. The pane} recognizes that it may be desirable to use only a single sensor when possible. At the very least, instruments such as the STS-2 should be utilized to extend the Tow frequency response to 0.01 Hz while still having adequate response in the short-period range. ISMS installations with both short-period arrays and a single broadband sensor should have some flexibility in the response criteria ofthe broadband system to ensure that low-frequency response is not sacrificed unnecessarily. The panel was very concerned to see that some of the updated working documents of the GSE specified low-frequency response down to only 0.04 Hz. This would greatly diminish the broader applications of the data and would even jeopardize the application of the ISMS data for routine functions such as computation of Ms. the surface wave magnitude, which is important for event discrimination. The pane! recommends that the low-frequency end of the nassband o f ISMS broadband systems extend to 0.003 Hz wherever possible. ~. - - There are some current applications for the seismic energy above 5 Hz in the earthquake monitoring and basic research communities, and the availability of globally distributed high-firequency data from quiet sites offers new potential for research on earthquakes and earth structure. Many current stations of the Uris, USGS, and FDSN networks readily can be (or have been) modified to achieve the high-frequency bandwidth of the ISMS specifications at relatively minor cost, and they will complement the primary stations. b. Seismometer Noise (10 dB Below Peterson's Low Earth Noise Mode! ([NATE. If this criterion is intended for the entire passband (0.02 to 20 Hz), it requires new instrument development, as we know of no broadband seismometer with a noise figure that is 10 dB below the ENM over the filet range. A combination of STS-! or STS-2 and GS- ~ 3 instrumentation can achieve this requirement over the passband 0.0 ~ to 10 Hz, and a combined KS540001 and GS-13 can achieve this requirement over the passband 0.03 to ~ O Hz. Certainly, this requirement, if achieved, is compatible with all applications of the seismic data. It would notseriously impact the research communi~'s current use of the data if the requirement were relaxed in the 10 to 20 Hz range to allow use of existing state-of-the-ar' equipment in fee SMS, and the pane! so recommends.

OCR for page 23
ISMS DATA CHARACTERISTICS 27 c Calibration (within S percent in amplitude and 5 degrees in phase) This is satisfactory for most uses of the ISMS data. It is satisfied by current IRIS, USGS, and FDSN stations. No change is recommended. d. Sample Rate 640 sps + 50 ms). This sample rate is suitable for most multiuse applications of ISMS data, as is the lower sampling rate of 20 sps as proposed in more recent GSE documents. The panel notes that in practice the 20-Hz high-frequency response objective is incompatible with the specified sample rate (40 sps). It may be desirable to oversample and filter to remove 50 to 60 Hz noise, and then resample to obtain the final desired high-frequency response. This could reduce the spectral density of the noise, which would benefit all applications of the short-period energy. Although the panel has not chosen to address the monitoring motivation behind the specification, some members noted that the preferred sampling rate is too low for research on and application of spectral characteristics to assist in discriminating between mining blasts and single explosions. Sampling intervals as much as four times smaller than the time between individual blasts in the mining explosions are needed, indicating sampling at up to ~ 00 sps. It is not currently realistic to achieve this sampling rate for continuous ISMS data at all stations; however, it may be desirable for certain stations, especially near mining areas. This high-sample-rate data could be saved on-site and accessed on demand. The pane! recommends that the sample rate be reevaluated e. Resolution (~S dB below Peterson's [N~ This requirement is driven by the desire to resolve very low amplitude signals that can be enhanced via stacking array elements. While this resolution can be achieved at existing FDSN stations using combined very broadband (VBB) and very short period (VSP) sensors, for isolated three-component broadband stations this level of resolution significantly departs from the needs of present seismological applications. The panel sees no advantage to digitizing deeply into the noise for single broadband stations. Furthermore, there are significant negative consequences. Large earthquakes at teleseismic distar~ces, and even moderate ones at regional distances, produce signals that exceed the finite dynamic range of a 24-bit system, leading to clipping if the system digitizes deeply into the noise. The pane! recommends that this requirement be relaxed for broadband t/'ree-component stations The pane' also recommends that the resolution levetbe tied to local noise. Since the [NM is rarely achieved, a more expeditious use of t/`e dynamic bandwidth wouldF be to base t/'efloor of t/'e resolution on a site's actual noise levels.

OCR for page 23
28 Comprehensive Test Ban Monitoring System Alternatively, the suggested CD noise resolution cart be achieved by the abolition of triggered broadband strong-motion sensors at the broadband stations. This can be implemented at low cost (as is a common practice at IRIS stations in seismically active regions), and it will ensure that ISMS data are not depleted in the very signals of greatest multipurpose use. f. Sensitivity f200 counts/nm at 3 Hz). There is no known nuantizer that can achieve the desired resolution over the entire passband with a sensitivity of 200 countsinm at 3 Hz. (The sensitivity would have to be set to 800 counts/nm to meet the resolution requirement at 20 Hz.) Sensitivity is adjustable depending on site noise, but for nuclear test discrimination at regional distances the critical passband appears to be in the range of 5 to g Hz. The pane! recommends specifying sensitivity goals at slightly Weigher frequencies or over a range offrequencies. g System Noise (10 dB below Peterson's LNM). The seismometer noise and sensitivity setting determines the ability to meet this requirement. While arrays can take advantage of low system noise to beat down natural background noise, this is not viable for individual fhree-component systems, so this requirement could be relaxedfor the latter. A more relevant reference point for system noise requirements is the local site noise, not the ENM. The pane! recommends that the system noise requirement should clearly specify the frequency band of importance. h. Dvnamic Range c~26 dB). The pane! interprets this design goal to apply to the digitizer capability. Existing widely used 24-bit quantizers achieve this range, so this requirement is compatible with multiuse applications of the data stream. However, if dynamic range is defined as the range from the ENM to the clip level, for the desired sensitivity (200 counts/nm at 3 Hz) this system will achieve no more than ~ ~ ~ dB at 20 Hz and 96 dB at ~ Hz. More recent GSE documents propose a more realistic 96 dB requirement. No specif c recommendation at this time. 1. Linearity ego dB over the passband). This is fully compatible with general seismic data requirements for diverse applications of the data. No change recommended.

OCR for page 23
ISMS DATA CHARACTERISTICS 29 i. Jiming.4ccuracv Al msecJ. This is Filly compatible with general seismic . data requirements for diverse applications of the data. No change recommended. k. Operating Temperature (-10 to 45C). A low-temperature requirement is probably needed oniv for certain sites and can be attained for most broadband systems with special insulating techniques and temperature controllers. Provided appropriate power is available, the proper environmental control system can increase the range of station locations, as is desirable for multiuse applications of the seismic data. No specif c recommendation at this time. Authentication Required. This is generally not needed for research applications. To the extent consistent with the monitoring goals, measures that are implemented should be such that they do not affect general use of the data stream. Some authentication procedures could involve significant modifications of existing instrumentation. The associated costs may limit the number of stations participating in the ISMS, which has negative implications for system performance. No specif c recommendation al this time. m. State of Health (at [east cilock statue. calibration status. ancivaultstatusJ. Such information is routine and desirable. No change recommencied. n. Format (one of the formats of the Group of Scientific Experts'. The broad international seismological community has established a standardized digital seismic data-exchange format Standard for the Exchange of Earthquake Data (SEED), which is now widely used in the FDSN. The pane! recommends that mechanism be established that would provide SMS data in SEED format in addition to other formats that might be used. O. Protocol (Telecommunication Protocol/Information Protocol (TCP/IP). This is compatible with other systems. No change recommended. p. Delav in Transmission (< 15 see). This is compatible with the needs of other systems. For global tsunami warning and earthquake hazard assessment, access within a few minutes is desirable. However, regional earthquake monitoring benefits from delay times of no more than a few seconds. No change recommended.

OCR for page 23
30 Comprehensive Test Ban Monitoring System q. Data Frame Length (<] sect. This is shorter than ir1 many existing stations, and it is not clear that such frames are an advantage. It is not needed for other uses of seismic data, which typically have data frame lengths of 2. ~ 5 to 8.6 sec for 20 sps and 1.075 to 4.3 see for 40 sps. Shorter frame lengths would negatively impact the data compression schemes used in many existing stations. More recent GSE documents have relaxed this requirement to < 60 sec. The panel recommends reconsideration of data frame lengtI, requirement. r. Data Access (Priority to International Data Center (ID C). then National Data Center (NDCJ). This item pertains to priority for communication with the station, not end-use distribution of the data. Provision of this access is technologically straightforward and can be implemented on existing systems such as those of the FDSN. it is very desirable that all data recorded at ISMS stations be made available promptly to the general research community. No specify c recommendation. s. Disk Buffer (7 days). This is readily achieved with current technology. No change recommended. t. Data Availabililv (>99 percent). This high percentage of reliability is driven by the needs of the nuclear monitoring function. This requirement diverges in practice from many other data acquisition systems because it is not cost effective. The panel expects that a 99 percent data availability requirement will lead to high operation and maintenance costs, thus limiting the funds available to support a large number of stations in the monitoring system, particularly in the auxiliary network. Even the 95% availability recommended for the auxiliary network in some recent GSE documents is likely to prove ur~duly restrictive. For example, the overall average IRIS network data availability is approximately 90 percent. The stated requirement could preclude use of these high-quality stations as part of the ISMS unless new funds are provided for the necessary level of maintenance. Academic and earthquake monitoring efforts typically prefer data from relatively dense and widely distributed networks, tolerating delayed access to some data and some gaps from individual stations, rather than the ISMS concept of data from a sparse network of highly reliable stations in near real time. With respect to auxiliary stations, tolerating reduced data availability, say at the 90 percent level, from an enhanced number of stations would provide data that would still achieve the overall desired availability and that would be of greater use to the seismological research community because of the expanded coverage. For example, given that the signal-to-noise ratio is such that two stations with 90% reliability have recordable ground motion, the probability that at least one of them will actually record is 99%.

OCR for page 23
ISMS DATA CHARACTERISTICS 31 Verification researchers also would benefit from the denser coverage and improved understanding of the regional geology and the wave propagation characteristics in interpreting data from an extended seismic network. The pane! recommends that separate and realistic data availability require- ments be establishedfor the primary and auxiliary networks. u. Timely Data Transmission (~98 percent). The availability of real-time data will not only benefit ISMS operations but will also extend the real-time data available to the USGS earthquake monitoring and basic research communities. Experience with four Global Test Seismic Network (GTSN) stations indicates a long-hauT communications link availability of 75 to 97 percent, depending on the station, so implementing this level of performance of real-time transmission for all primary stations will be very challenging. No specif c recommendation at this time. v. Station Location (within 100 m. array elements within ~ m relative J. This is a routine requirement, although a specific reference frame for location, precision, and accuracy should be given. No specif c recommendation. w. Seismometer Orientation (known within ~ degree). This is an extraordi- narily high accuracy, not routinely achieved with any borehole instrument (KS360001 and 54000T orientation is +/- 3 degrees), and vault-type instruments can be oriented this accurately only if a suitably accurate survey mark is provided in the vault. More recent GSE documents suggest that 3 is an acceptable specification. The pane! recommends relaxation of the orientation tolerance. Desired Raw and Processed ISMS Data Streams This issue watt be taken up in detail in the next chapter, but some initial response is warranted in the context of the first charge. There is wide enthusiasm for timely and straightforward access to the broadband three-component data streams from the (continuous) primary and (segmented) auxiliary stations of the ISMS, as well as to continuous data from the auxiliary stations, which will not be collected by the ISMS. The continuous primary data can be directly incorporated into real-time analyses conducted by the USGS global earthquake monitoring system, complementing data collected through other networks. In addition, the broad seismological research community has numerous applications of real-time seismic data analysis, and timely

OCR for page 23
32 Comprehensive Test Ban Monitoring System access to the ISMS data would enable maximum utilization of the data in diverse applications. Segmented auxiliary station data, which will be acquired within several hours of events of interest, also will be valuable for rapid analysis by the USGS and basic research communities. Specific suggestions as to how to ensure rapid and full access to the complete waveforms from the broaciband sensors in the primary array, as well as all data from the auxiliary network, will be provided in the next chapter. The ISMS processing will include automated and analyst-reviewed measure- ments of phase arrival times, slowness measurement from the primary arrays, and array beams, f-k spectra, and event locations. The USGS indicates that the phase and Ut/~^ measurements and bulletins from the TDC, both automatic and reviewed, would be useful, but not essential, resources for their operational program. The USGS also indicates it is unlikely that it will have serious use for products such as f-k spectra in the near term. USGS operations do not currently place a high emphasis on global array data, either raw data or formed beams, but this situation could change in the future. There is interest from the general research community in obtaining short-perioc} data from the primary arrays. The primary interest is in the original array data for events of mb > 4 for the several-minute time window encompassing teleseismic phases. Access to data from the individual array elements, rather than stacked signals, is likely to be of interest to members of the research community, particularly with respect to discrimination and deep earth structure research. However, the vast quantity of data involved is such that it would be costly to duplicate the archive of the full data set. Developing a procedure for accessing the ISMS data archive for tailored user requests of array data seems to be the most attractive option. There does not appear to be a general requirement in the research community for intermediate products such as the ~ 00 continuous beams formed by each array or f-k spectra as long as access to the raw array data is established by some convenient procedure. This chapter has considered data characteristics proposed by the GSE for the TSMS seismic stations in light of research requirements both for general seismology and for earthquake and nuclear test monitoring and detection. The general objective of the panel's recommendations is to ensure that the TSMS seismic data are as inclusive and as broadly applicable as possible. The pane! has therefore recommended some changes in Tow-noise level requirements, sample rate requirements, sensitivity goals at higher frequencies, and data frame length. The pane] also recommends adopting methods for ensuring rapid and full access to data streams. These are discussed at greater length in the next chapter. This chapter has dealt with the data characteristics of the proposed permanent ISMS monitoring stations. Characteristics of portable instruments have not been covered, but the pane] notes that use of appropriate portable apparatus would increase coverage temporarily in an area of particular interest.