Golfballs on the Moor: Building the Fylingdales Ballistic Missile Early Warning System
Download 134.75 Kb.
|
BMEWS OperationIn contrast to the delays with Project Legate, the BMEWS radars themselves were ready on schedule, providing data for transmission to NORAD. Fylingdales became operational on 15th September 1963, and transfer to RAF operational control was on 15th January 1964. Experience with early warning messages came earlier, however, as a tele-type link was set up as soon as the Thule BMEWS went operational in October 1960. This link, provided by the Americans, initially supplied messages from NORAD to Fighter Command ADOC, much to the annoyance of Bomber Command, which insisted on simultaneous access to BMEWS information. There was some concern over the interpretation of these messages because the alarms generated by BMEWS were based on computer analysis of potentially ambiguous data: ‘BMEWS does not provide clear-cut positive warnings. Instead it indicates degrees of probability that missiles are on the way, which have to be interpreted. So considerable are the difficulties of interpretation that VCAS decided that D of Ops (AD) and a small team should visit the States in November 1960 to discuss the problem.’57 The Thule and Clear BMEWS used fixed radars that scanned across two elevations, creating what were known as ‘fans’ of coverage. Signals received by the lower fan of the radar were ‘immediately passed through a series of filters intended to discard reports of objects that are clearly non-dangerous’, such as satellites, meteorites, and planets.58 The six checks designed to eliminate objects whose trajectory and velocity meant that they would not fall within the target area were: (a) Range-Change Rate Test. This compares the value of the target’s range-change rate with a threshold value that is a function of its range and azimuth. The assumption is that an object with a range-change rate above a pre-calculated threshold, for any given point in space, will not return to Earth. (b) Azimuth-Change Rate Polarity Test. This determines the polarity of the target’s observed azimuth-change rate as a function of azimuth. For given azimuth angles and particular azimuth-change rate polarities, no impact can occur in the area to be warned. (c) Orbital Plane Test. This determines whether the orbital plane of a target trajectory falls outside the limiting planes containing the area to be warned. (d) Energy Test. The energy equation of the target is solved to check whether the energy is greater than the escape energy, thereby rejecting high energy objects such as space probes and meteors. (e) Perigee Test. This test determines the magnitude of target’s perigee and rejects objects whose perigee is greater than the Earth’s radius. (f) Ground Range Test. The final test ascertains indirectly whether the range over the Earth’s surface covered by the target in its flight will exceed a certain value, in which case it rejects it.59 However, there remained some ambiguity in how remaining threats were assessed: ‘All reports not discarded by these checks are considered potentially dangerous objects and each is given a confidence value by the electronic processing computer. The detailed means of assigning confidence values to reports is complicated and wide open to change in the light of experience. Many factors affect it: for example, the presence of high background noise would tend to reduce the final confidence value assigned, and an object that was subjected to checking by the upper fan (“two-fan correlation”) and survived it would thereby receive a greatly enhanced confidence value.’ 60 ‘Threat summary messages’ were then prepared by the computer, and forwarded to NORAD, on the basis of cumulative numbers of reports over three time periods. These counted ‘the total of confidence values currently accumulated in each of three “windows” – one window summing the preceding 5-minute period, a second the preceding 15-minute period, and the third the preceding 45-minute period.’ In addition, the computer held ‘lower-fan data concerning potentially dangerous objects in store for correlation with upper-fan data’. It took about 210 seconds for an object at long range to pass from the lower to the upper fan of the AN/FPS-50 radar. If a tracking radar was available on site, the computer would arrange for it to acquire and track each object in turn. ‘From the fan-to-fan correlation and/or tracking radar data the computer then works out the trajectory of each potentially dangerous object and forwards launch and impact point prediction data (place and time) to NORAD if the impact point lies within the area to be warned.’ 61 Of some concern was the fact that the Thule BMEWS had already sent the highest level of alarm to NORAD: ‘the Americans frankly admitted that an incident had occurred in October, during the first weeks of the operation of the Station. Apparently the lower beam of the radar recorded a heavy missile attack, which the computer duly passed and recorded as a warning state. The Americans took no action on this warning which was subsequently cancelled some few minutes later when no tracks were reported through the upper beam of the radar. The false report was caused by the moon. The Americans had thought that they have eliminated the effect of the moon, when processed through the computer, but admitted that something had gone wrong with their calculations. One answer to the problem, if they can do it, is to get the computer to eliminate any such reactions in future; and to supplement this the Americans have ordered urgently a tracking radar which establishes whether the reports are actually due to missiles in flight or some other cause’62 In the meantime it was appreciated that BMEWS data should be treated with some scepticism: It must be clearly recognised that the BMEW system is initially on a very restricted operational capability, in fact it is understood from the USAF that their evaluation of the system may well take some 6-9 months before the precise quality of generated information can be assessed. It must also be recognised that false alarms are possible especially during the initial period and the alarm level figure must be interpreted not only in conjunction with the impact prediction but with all other forms of strategic and political intelligence. For these reasons the point has been made by NORAD that high defence alert states will not be ordered solely on receipt of BMEWS information until such time as the system has been fully evaluated.63 In particular, there was no question that BMEWS warning alone would at this stage trigger any military response: ‘When asked by a direct question whether NORAD would issue the highest defence alert on BMEWS warning alone, General McColpin answered with an unqualified “No”.’ 64 In fact, ‘NORAD had not yet issued a detailed SOP [standard operating procedure] on action to be taken on receipt of a BMEWS alarm’ and ‘it appeared that lower alarm levels than 5 would result, in NORAD, in nothing more than the alerting of personnel.’65 Once produced, US BMEWS warning data was supplied to the UK via a teletype circuit. The format for missile warning messages would have the words ‘BMEWS warning’ followed by: the alarm level (a number 3, 4 or 5 with 5 indicating the greatest threat); predicted impacts (as 0, 1, 2-5 or above 5); and the estimated raid size (low, medium or high, with low indicating 0-5 missiles, medium 6-15 missiles, and high more than 15 missiles).66 However, by the following year the Americans had changed both their method of calculating alarm levels and the way that they would be displayed. For Thule a ‘points scoring system is in force with 1 point for a single-fan event and 10 for a two-fan event, but not more than twenty points can be scored from single-fan events. Three alarm levels are used, 3 for 40 points, 2 for 60 points, and 1 for 80.’ At Clear an ‘identical system is used but, because of lack of operating experience, the points required for an alarm level are doubled. It is expected that by the end of the year the same values will be given to Clear as Thule.’ To get a system alarm ‘the computer adds together the points from each site, giving appropriate weight to each. There is a possibility of double-counting where both sites register the same object but since in this case two different system will have accepted the object as a missile, the extra weight given to it is acceptable.’67 This change in way US data were displayed led to further friction between Bomber Command and Fighter Command when the former was not informed of the change. As BMEWS was considered an element of the UK’s air defence system it had been seen as appropriate for Fighter Command to take charge of Fylingdales. However, it was the V-bombers that would have to react most rapidly to an early warning message. This issue had already been aired in 1960 when it was made clear that: ‘It is essential that information drawn from the BMEW System be displayed in the Bomber Command Operations Centre in the same time scale as at the Air Ministry Operations Centre or Air Defence Operations Centre. This information must not be delayed in any way or filtered and ‘re-told’ through any other organisation.’68 Now, a year later, Bomber Command was not impressed that it had not been informed of the display change: The display has been changed. It now shows Alarm Level (combined from both sites), number of events in last 5 minutes (an event is a response with missile characteristics), number of impacts predicted in US and time to go before the first impact. Maps show predicted impact areas and launch sites. … The information which will be passed to ADOC and BCOC is alarm level, number of impacts predicted in US, time to go before first impact. Formerly, it was alarm level, raid size and predicted impacts. NORAD claim that these changes were signalled to Air Ministry, copy to ADOC last June, but the information was not received at BCOC. This failure of liaison might have caused disastrous confusion in an emergency. In addition it has caused the loss of several hundred pounds spent on a display system which is now useless.69 Further modifications to the US reporting system were summarised in February 1963: Radar reports were not ‘classified as low confidence and high confidence reports. Low confidence reports are those produced from radar returns detected in a single radar fan and may arise from space objects other than threatening missiles and even from the various forms of interference that are experienced. A high confidence report is one which results from objects detected by 2 fans of the same radar and subjected to more stringent tests than are low confidence reports, to determine that a detected object has all the characteristics of a threatening ICBM.’ All reports were assigned a score, 1 for a low confidence report and 11 for a high confidence report increasing to 15 if the high confidence report had been further substantiated as an ICBM by tracking. Running totals of scores over periods of 5, 15 and 45 minutes were then applied to alarm level thresholds in a Display computer and the highest threshold exceeded determined the alarm level figure generated and displayed. The purpose of the different periods over which the running totals were accumulated was ‘to ensure that raids comprising missiles fired singly with comparatively long periods between firings may still generate the highest level of alarm.’ Low confidence reports alone could not generate an alarm as they were limited to contributing a maximum of 10 to the running totals. Thus, ‘at least one high confidence report, together with the maximum number of low confidence reports is required to generate the lowest level of alarm.’ Moreover, ‘every missile detected in both upper and lower beams (fans) will immediately produce a high confidence report. It is therefore possible and in fact probable with the type of raid expected, that sufficient high confidence reports will simultaneously be generated to trigger off the highest level of alarm as the first warning of attack.’70 Of course, understanding the US approach to BMEWS reporting, while obviously useful, did not resolve all the issues regarding the UK approach. One issue that was quickly decided was to what extent threat assessment should be computerised. A November 1962 memo noted that: ‘The formula for computation of the BMEWS UK alarm level will soon need to be defined. RRE have asked whether one computer programme should be used for computing the UK alarm levels or whether we should have a number of programmes which would take into account other “intelligence”, such as the political situation at any specific time, the ECM level, equipment serviceability and other pertinent factors. I consider that the BMEWS UK alarm level should be based purely on the radar returns and resultant MIP computer threat summary outputs to our display computer. This would require one programme and would provide a standard RAF/USAF method of generating BMEWS alarm levels. Assessment of the overall military situation would then more correctly be left to ADOC and BCOC staffs, who would view the BMEWS alarm levels in relation to all the other forms of “intelligence” available to them. I think this assessment is essentially a human responsibility – not a computer operation.’71 What did need to be computerised was the operation of the radars, and in this regard the US experience was not directly applicable to the UK’s use of Fylingdales. Both the type of radar involved and the military requirements of Fylingdales differed somewhat from the first two BMEWS. Whereas BMEWS Sites 1 and 2 scanned constantly at two elevations, producing two ‘fans’ of coverage, Fylingdales was originally conceived as scanning only one elevation. Further confirmation of threats and more accurate trajectory information required the use of the third ‘lock-on’ tracking radar: The radars at Site 3 are tracking radars with a single beam. It is considered at this time that reports produced by missiles passing through a beam in the scanning mode will be low confidence reports. To produce a high confidence report it will be necessary to lock-on and track individual missiles. Low confidence reports will be produced by scan information and passed to the lock-on radar for individual attention. 72 However, in normal operation there was only one radar allocated to the tracking function, and so there were issues about how long it would take to lock-on to each suspected threat: From the standard rest position of 75 degrees in azimuth, the lock-on track radar will take a maximum of 7 seconds to reach any point in the radar cover, but it is more likely to be 2 or 3 seconds since the main IRBM threat is from East Germany and nearby Soviet territories. The radar antenna and electronic equipment requires 2 to 3 seconds to settle down after rotation and actual average tracking time will be 5 to 10 seconds. Thus, one high confidence report can be produced in a period of 10 to 20 seconds after generation of scan data. 73 In addition, there was also the question of how to allocate priority between ICBM threats, which were mainly of concern to the US, and IRBM threats, which were the primary threat to the UK: ‘The doctrine governing allocation of the lock-on track radar to suspected IRBM or ICBM reports favours IRBM tracking in a ratio approximately 2 to 1, with initial allocation in a mixed raid to IRBMs. Low confidence scan information on any number of missiles can be produced in a period of from 7 to 30 seconds after initial entry of missiles into the scan cover of the radar. Thus, the first high confidence report may be produced on an IRBM in anything between 17 to 50 seconds with a second high confidence report, assuming more than one missile, in a further 10 to 25 seconds. If ICBMs are present in the raid, the lock-on/track radar will then switch on an ICBM for approximately 20 seconds then return to IRBMs. If no ICBM is present in a raid, the lock-on/track radar will continue processing IRBMs.’ 74 A major concern for the British use of BMEWS was the fact that the shorter distances involved meant that warning times might be very limited. To ameliorate this concern the lock-on radar would be directed towards the direction from which IRBMs would be expected to arrive in the shortest time: Bomber Command does not maintain a continuously airborne retaliatory force and maximum warning time of IRBM attack is required. Reference to realistic trajectories for the worst case of IRBM attack, (15 degree re-entry angle from 450 mile range) shows that low confidence reports will be produced with 3.5 minutes to go before impact. Because these missiles would be in a sector which requires very little movement of the lock-on/track radar from rest position to target, two high confidence reports should be produced with at least 3 minutes to go before impact. This is the worst case, however, the alarm level must be adjusted to this.75 The UK Threat Summary was ‘calculated solely from information obtained from Fylingdales and … concerned with IRBMs aimed at the United Kingdom’. It comprised: ‘(1) The Alarm Level. (2) The threat index, the percentage of threat activity towards the next Alarm Level. (3) The number of missiles that have been tracked. (4) The minutes to impact of the “earliest to arrive” missile.’76 The US Alarm Level was calculated from all 3 BMEWS and was concerned solely with ICBMs aimed at the North American continent. This information was presented on the BMEWS displays at: The Air Force Operations Room at the MOD, the Bomber Command Operations Room at Headquarters Bomber Command, the Standby Bomber Operations Room at Headquarters No 1 Group, the Air Defence Operations Centre at Headquarters Fighter Command (which would move to West Drayton in the Linesman era) and RAF Fylingdales. Following deployment of Polaris there would also be a display at Northwood. Directory: esrc-files -> outputs outputs -> Hunter, Helper, Leader, Thief? Unemployed Youth in the Indian Himalayas a report on some preliminary fieldwork findings Jane Dyson and Craig Jeffrey, 7 Download 134.75 Kb. Share with your friends:
|