The operational and economic responsibilities for the individual stations should stay with the owner countries. They could also be discharged in bilateral co-operation between countries. In this connexion the notion of "sister observatories" and other organizational matters have been taken up in an earlier Swedish working paper (CCD/405). It is important that the individual stations of a network have comparable, if not equal, capabilities. A station with an exceptional sensitivity could yield detections which could not be verified by other stations. Similarly a relatively unsensitive station would make little contribution to the joint capability of a network.

The number of stations in a network could be kept fairly small. It has been demonstrated for example during the International Seismic Month in 1972 (CCD/388), that carefully analysed data from appropriately selected stations can provide detection and location capabilities comparable with and even superior to those routinely achieved at present from the analysis of data from all seismological stations on the globe. A limited number of stations will also keep the data flow at a manageable level.

A global seismic station network might be supplemented by networks of local stations, to monitor regions of special interest, e.g. areas containing thick deposits of low coupling media.

It is important that careful analysis is carried out at each individual station. Besides arrival times, signal amplitudes and period data, which are generally reported at present, the analysis should extract special identification parameters, based on short as well as long period recordings and also information on depth phases and other secondary phases. The identification parameters could well differ from station to station depending on the facilities available and on the receiving conditions for seismic signals. The analysis at the stations would materially reduce the amount of data to be transmitted for joint analysis at the data centre described later on in this paper. Today only a few stations routinely conduct the type of analysis outlined here.

The operation of several of the arrays are at present in a more or less experimental stage. In a monitoring situation one has to put strong demands on maintenance and calibration, to obtain high quality data and to enhance the operational reliability of the stations.

2.2 An example of a network

The map in Figure 1 shows the geographical distribution of the seismic stations selected here to constitute an example of a highly sensitive seismic station network having global coverage. Networks of local stations, which might be desirable, are not discussed here. The stations, which are listed in Table 1, can be grouped into the following categories according to their equipment:

Ibid., 1972, pp. 590–611.

This makes in all 46 stations, 37 of which have short period and 25 long period instruments. The stations are distributed over 26 countries, as can be seen from Table 1. The total number of stations is somewhat larger than the minimum necessary in order not to make the network critically dependent on certain stations. The capabilities of this network would not be significantly reduced even if some ten stations became inoperative for certain events or ceased to operate in the network.

The station network in Figure 1 has been selected to obtain a global coverage utilizing as many stations as possible already existing or under installation. Only two African stations have been selected. The extremely favourable recording conditions in Africa would, however, make it worthwhile to consider the establishment of a few more well equipped stations especially in the Central and Northern parts of the continent.

3. Data exchange

Parameters extracted from the recordings at the individual stations should be transmitted regularly in the form of a bulletin to the data centre. The total number of parameters in the station bulletins is estimated to be not more than 20 per event. A highly sensitive array with both short and long period recording systems can be expected to detect roughly 30 events per day. This would mean that the average number of parameters to be transmitted from a station is not greater than 600 per day. It is important that every station transmits the data in a consistent format, but it is not necessary that all stations use the same format.

Ordinary telex channels provide a convenient way of transmitting seismic bulletin data. The global telecommunication system of the World Meteorological Organization has been discussed for transmission of seismological data for this and for other purposes. These systems are simple and easily accessible to most countries and are also compatible with computers. Exchange of full recordings, which is supposed to be conducted on a much smaller scale, could probably be handled by ordinary mail. If recordings are to be sent as digital magnetic tapes, the problems with achieving compatibility between computers must be considered. Large and technically advanced computer communication networks like the ARPA NET (CCD/388) would

be helpful but not necessary for handling the limited amount of data discussed here.

4. Data centre

The task of the international data centre would be to assist States wishing to monitor the CTB, by collecting and processing data about seismic events. The data and the results arrived at would be made. available to the parties. In this way the data centre would be particularly valuable to countries with only limited expertise and facilities. The political assessment of the events analysed by the data centre could be left to the parties of the treaty.

In particular the data centre should have the following main functions:

-Collect data from the network of selected seismological stations; -Define seismic events and estimate their epicentres and depths; -Apply various seismic identification criteria to the events defined; -Distribute the results to interested parties;

-Conduct consultations and make inquiries with designated institutes to obtain additional data and information about events insufficiently described by the data routinely available; and

-Provide experts to observe, by invitation, nuclear explosions for peaceful purposes and large chemical explosions.

The detailed procedures of the centre should be further discussed. The procedures outlined here have been worked out to facilitate and promote such discussions.

The procedure of analysing seismic events, which could be undertaken in various ways, would have to be agreed upon among the parties concerned. Even if the current seismic event identification techniques are highly effective there will probably be a small residue of events defined by the data centre which could not be sufficiently analysed by routinely applied teleseismic methods. Further analysis would then be made by acquiring additional seismological data, e.g. from local seismic station networks or by applying more refined analysis methods than would be applied on a routine basis. Satellite photographs for relevant areas could also be requested and analysed in order to facilitate the clarification of residual events.

If the nature of a seismic event can not be confidently clarified, the centre could consult designated agencies in the country, to which the event had been located, for further clarification. If doubts still remain about the nature of the event, the countries concerned should be notified and they could then make their own interpretation of such events and take the steps they may find appropriate. Obviously, the data centre should confine itself to scientific tasks. Political evaluations of the events should be left to the Governments parties of the treaty.

The monitoring system should operate with a time delay that would be sufficient in order to take appropriate political action after a violation of the treaty. The length of this time delay would have to be agreed upon.

The observation, by invitation, of nuclear explosions for peaceful purposes and of large chemical explosions would be an important element in the monitoring of a CTB. The participation in such activities would be an important task for the centre. The details should, however, be analysed separately.

To conduct the tasks outlined above, a staff of some 40 professionals and technicians and appropriate technical equipment and computers would be required for the data centre. The estimated cost of such a data centre is of the order of $2 million per year.

The data centre could be set up as an independent international institute or be included in an already existing international agency or part of a disarmament monitoring agency of the kind earlier proposed by the Swedish delegation (CCD/PV.610).

5. Expert study of co-operative international measures

To promote a comprehensive discussion of co-operative international measures to facilitate the monitoring of a CTB, a group of experts should be appointed to investigate this matter further. Such an expert study should include discussions of the following items:

-A suitable global network of seismological stations, based primarily on existing and planned installations.

-The data analysis required at the stations to produce data to facilitate the detection, location and identification of seismic events. -The data transmission facilities for the exchange of data between the seismological stations and the data centre.

-The facilities and procedures needed at the data centre to detect, define and locate seismic events from reported data.

-The acquisition by the centre of satellite data and other nonseismic data.

-The methods, seismic and non-seismic, and the procedures to be used at the data centre for the analysis of observed events.

Such an expert group should be appointed without further delay and should report at regular intervals to the CCD and should deliver its first report as soon as possible.

The matters of observation, by invitation, of nuclear explosions for peaceful purposes and of large chemical explosions are also of great importance but would have to be investigated separately.