section was made. A silhouette of the F-4B was then
- 10 - Wreckage portions of the DC-9 forward fuselage were removed, and a full-scale three-dimensional mockup of this section was made. A silhouette of the F-4B was then constructed with lumber. This silhouette was placed in various positions and attitudes to attempt to match the two distinct damage paths through the DC-9. one path was long and narrow, oriented at an angle of 300 to the ~C-9 fuselage reference plane, and passed through the fuselage in the area below the main passenger loading door and first 10 windows.
The other path was rectangular and passed through the forward, lower, cockpit area. When the F-4B vertical stabilizer was positioned at approximately FS110 (the fuselage area beneath the windscreen) on the left side of the DC-9 the right wing was in the damage area under the cabin door and windows. Both damage paths were on a descending angle of approximately 200 through the DC-9c However, both damage swaths were larger than the F-4B structure, and this angle could vary as much as 100 in either direction. It is not known whether the downward trajectory of the t~o swaths resulted primarily from the relative flightpaths of the two aircraft or the progressive disintegration of the F-4B structure as it passed through the DC-9.
1.13 Fire No evidence of in-flight fire was found on the oc-9; however, the F-4B caught fire following the collision.
There was a severe ground fire at each of the main crash sites. A total of seven fire trucks, two helicopters, one air traffic controi unit, and approximately 72 officers and men responded to the fire alarm.
1.14 survival Aspects This was a nonsurvivable accident for the occupants of the DC-9.
The midair collision was survivable for the occupants of the F-4B. The RIO successfully ejected and he was subsequently rescued uninjured. The pilot was not able to eject and the F-4B collision with the ground was non- survivable.
BuNo458 was equipped with a Martin-Baker B7 rocket ejection seat in each cockpit. This seat was not designed to be fired through the canopy, and incorporated a canopy - 11 - interrupter block in the actuation linkage to prevent such an occurrence.
The front seat face curtain, the primary means of firing the seat, was not recovered. The alternate firing handle had been actuated. Distortion of the actuation linkage indicated that the canopy interrupter block prevented further movement of the mechanism and subsequent ejection of the pilot.
In addition to the canopy unlock system provided in the ejection sequence, two manually operated systems are provided. However, regardless of which method is used to unlock the canopy, cases have been reported wherein the front canopy failed to jettison when the aft canopy was jettisoned first.
As a result of these occurrences a change in the canopy jettison mechanism was instituted to incorporate ballistic canopy thrusters to assure that the canopy would separate from the aircraft. This modification had not been installed in BuNo458. This change was being incorporated in all F-4B aircraft on a fleetwide basis, and modification of aircraft at MCAS El Toro was scheduled to begin in July 1971.
1.15 Tests and Research A radar f ligbt check of the San Pedro radar was conducted on June 8, 1971, using an F-4B.- Routine scheduled maintenance had been performed on the system between the time of the accident and the flight check. The RIO bad not been formally interviewed by safety Board investigators at that time. Consequently, the flight track was only an approximation of the presumed track. The San Fedro system was capable of tracking the primary target of the F-4B above 7,500 feet. several controllers commented that they had never seen the radar perform so well.
The Safety Board coordinated with the Federal Aviation
Administration (FAA) and the u. s. Marine corps to conduct
another flight check of the San Pedro radar on June 16, 19710 The track of BuNo458 described by the RIO was duplicated as closely as possible on three runs, with some variations in the altitude on one run. Three additional runs were conducted in the general area the flight traversed, but with flight track and altitude variations as suggested by the witness group. The radarscope, channel; - 12 - and control settings were the same as at the time of the accident, except that the secondary target was offset so that it would not interfere with tracking the F-4B primary target. Tracking continuity was poor and the primary target was visible less than 50 percent of the time. The secondary target did provide assistance in following the aircraft movement during voids in the primary target coverage.
Photographs of the test runs, as displayed on a maintenance monitor, were taken on virtually every sweep of the antenna.
These photographs were studied by the air traffic control group. The examination corroborated the initial reaction to the tracking quality of the F-4B target, but it also demonstrated that the primary target alone was not of sufficient strength to assure notice by a controller who was unaware of the aircraft presence. · The F-4B in each test was not configured the same as BuNo458 at the time of the accident. The first test F-4B was in a clean configuration, and the second F-4B test aircraft was equipped with two large wing tanks. A baggage tank was installed on the fuselage centerline of BuNo458.
Consequently, BuNo458 represented a larger reflective cross- section for radar detection than the first test aircraft, and less reflective surface than the second aircraft. The validity of the flight checks also was compromised by such variables as meteorological phenomena and deterioration in operating parts of the radar or improved performance due to replacement of failed parts.
A visibility study was conducted to determine the physical limitations to vision from the cock~it of each aircraft. A flightpath for each aircraft was reconstructed.
(See Attachment 1.) The collision geometry and closure rates for the last 40 seconds also was reconstructed. (See Attachment 2.) The flightpath of RW706 was based on the flight recorder data, and the F-4B flightpath was predicated on the statement of the RIO. A dual lens camera was used to record a panoramic view from the design eye-reference point at each crewmember•s station. (See Attaqhments 3 and 4.)
These binocular photographs show the position of each air- craft in the field of vision of each crewmember, based on his fixed-eye-reference point. Naturally, any movement from this position would affect the location of the other aircraft in his field of vision.
- 13 - In the course of thi$ investigation, the McDonnell- Douglas Corporation provided information concerning roll and pitch rates for the F-4B aircraft. The following bas been extracted from the supplied data: 1/4 Full Stick 1/2 Stick Throw Subject Throw Stick Throw What would be the maxi- mum rate of steady nose- down pitch? · -10 deg./sec. -4 deg./sec. -2.5 deg./sec.
Concerning the nose- down pitch rate, what would be the time interval required from the first control input to; Achieve initial air- (a) 0.10 sec. 0.10 sec. 0.10 sec.
craft movement?
(b) Achieve the maximum steady nosedown 6.0 sec.
1.0 sec. 6.5 sec.
pitch rate How many degrees nose- down would have been achieved at the point that the maximum pitch 15.0 deg. 9.0 deg. 5.0 deg.
rate had been attained Altitude lost and air- speed at 200 nosedown 590 ft.
pitch: a) alt. loss 16 ft. 210 ft.
b) KTAS 420 426 Ci) 200 loss and airspeed Altitude flightpath angle: nosedown a) alt. loss 385 ft. 620 ft.
180 ft.
KTAS 424 430 437 b) In addition to the above, the data indicated that a bank of 300 could be achieved in as little time as 0.75 second. If roll-and-pitch control inputs are coupled during the maneuver, the time to achieve a given bank/pitch - 14 - ~ttitude is less than the time required if the attitude is achieved as the result of two separate maneuvers. In this instance, the data indicate that a 200 nosedown, 300 left bank attitude could be achieved in less than 3 seconds.
1.16 Other The Staff Vice President, Flight Operations for Air west, stated that attempts were made to foster crew vigilance and scanning by minimizing crew duties in the cockpit, use of checklist procedures, encouraging use of the autopilot as much as possible, and through emphasis in their training program. However, tbe various manuals and training programs did not specifically contain any statement relating to lookout doctrine or scanning techniques, nor did the company have any plan to implement such a program.
The pilot of BuNo458 received training in lookout doctrine and scanning techniques in flight school. After assignment to the squadron, the pilot and RIO received additional training on scan techniques with emphasis on tactical intercept and pursuit. Each pilot and RIO receives formal upgrading and refresher vision training at least once every 3 years. Additionally, the crews routinely include reminders in lookout doctrine during briefing for each multiple aircraft flight.
2. ANALYSIS AND CONCLUSIONS 2.1 Analysis (a) ATC and Radar Factors The primary function of radar is to ~rovide the controller with a visual presentation which will assist him in the control and separation of known traffic. It also provides the controller with another limited capability that of ~roviding separation of identified from unidentified traffic through the medium of the traffic advisory when workload permits. In some cases, due to the technical limitations 0£ the radar equipment, adequate separation has not been achieved. Because of the mix of known and unknown traffic it is not only incumbent upon aircrews to maintain a high degree of vigilance to "see and avoid", tut also upon the controllers in monitoring the radar display. In this accident three independent radar systems failed to detect the primary target of EuNo458 and as a result no warning was - 15 - given to the crew of RW706 regarding the direction and distance of the hazard. If the crew of Rw706 had been provided with this information their chances of seeing and avoiding the other aircraft would have been enhanced. One solution to the limitations of radar would be the establishment of some minimum standard of reflective capability for all aircraft and the incorporation of some form of signal enhancement equipment aboard all aircraft, as necessary to meet the standard, as previously recommended by the Board. (See Report Number: NTSE-AAS-70-2, pages 119- 128.)
The radar coverage chart (classified for military security) for the San Pedro system indicates that the collision occurred at an altitude which is within the basic radar line of sight coverage. The limitations to radar advisory service within that coverage area include more than the controller workload. Other factors affecting detection of primary targets include: (1) Radar cross-section presented by the design and configuration of the aircraft (2) weather conditions such as precipitation and temperature inversions (3) Ground clutter (4) Elind spots In this instance, detection of BuNo458 was hampered by the aircraft radar cross-section and a temperature · inversion. Although simulations of the flightpath indicate that the primary target was intermittently detectable, the low probability of such detection is dramatized in the following computation: The total elapsed time a target would have been detectable was 120 seconds, equal to 10 sweeps of the antenna. At approximately 420 knots (7 miles per minute), the aircraft would travel 1.4 miles during each sweep. The target would actually move a total of 2.5 inches, or 0.25-inch/sweep, across the 21-inch display. The small time element involved and short 9istance moved, in combination with the prcbability of less than 50 percent primary target tracking continuity, indicate that it would have been extremely difficult for the controllers to differentiate between normal clutter and an aircraft return, if any target was displayed at all.
- 16 - The volume of traffic and controller workload associated with the R-18 sector were sufficiently light to permit radar traffic advisories if requested. Advisories on possible conflicting traffic were being given to other controlled aircraft during the time period surrounding the collision.
All four controllers associated with the activity at the position stated that no primary targets were observed in the vicinity of RW706. Consequently, the Board concludes that no readily discernible target from EuNo458 was displayed.
If a request for radar advisories had alerted the controllers to the presence of an aircraft in that area, any intermittent or questionable target sighted could have been tentatively identified as EuNo458. The R-18 controller could have advised RW706 of the conflicting traffic under these circumstances.
(b) Reports of Aircraft Acrobatics During the investigation considerable public attention was focused on witness reports of an aircraft performing acrobatics in the vicinity of the collision. The RIO testified that only one aileron roll was performed by the pilot of BuNo458, as he leveled off at 15,500 feet. An analysis of the flight from NAAS Fallon indicates that there was insufficient time available for any repeated maneuvers to have been performed. The witnesses might have been observing another aircraft, or they were actually viewing the gyrations of BuNo458 following the collisicn. Whereas no specific Federal Aviation Regulation prohibited the aileron roll, the ability of the crew to see other aircraft during the maneuver was unquestionably minimal due to the rapidly changing attitude and the acceleration forces imposed. The Board concludes that the aileron roll had no other significance to the accident, since the two aircraft were separated by approximately 13 miles at the time.
However, it was imprudent of the pilot to perform such a maneuver in other than an acrobatic area.
(c) Operational Factors This accident is another example of a heterogeneous mix of VFR and IFR traffic, with each aircraft ccmplying with applicable regulations, resulting in a midair collision.
several factors in the operation of the two aircraft combined to provide the conditions suitable for a midair collision ..
- 17 - 1. Operation of BuNo458 Mechanical difficulties with BuNo458, and the resulting operational decisions, ~laced the aircraft at low altitude and high airspeed, instead of in the APC, as would normally be expected on cross-country flights. The transponder had failed on the previous day, making entry into the positive control airspace dependent on the discretion of the air traffic control facility. When the oxygen system also became defective, with no opportunity to repair either system, the decision to proceed to MCAS El Toro at relatively low altitude was the obvious solution to both problems. The transponder was not required, and cockpit pressurization negated the physical need for supplemental oxygen, even if the leak depleted the entire supply. The oxyqen leak did increase, and most of the flight to NAAS Fallon was flown without supplemental oxygen. At this point, the pilot was instructed by higher authority to complete the flight with the defective systems. An additional significant factor in the operation of BuNo458 was the high-cruising airspeed, which is typical of modern jet aircraft. The high airs~eed was used to avoid high specific fuel consum~tion and the less stable flight regime encountered at slower airspeeds. consequently, the proba- bility of visual detection was minimized by the speed, size, and unexpected presence of BuNo458. · In addition to these operational constraints imposed on the pilot of BuNo458, consideration of traffic and weather conditions was also evidenced in the planning and conduct of the flight ..
The pilot of BuNo458 was sufficiently aware of the heavy volume of traffic in the Los Angeles area to alter his flight to the east in order to avoid any conflict. He further demonstrated concern for adequate vigilance by climbing to 15,500 feet because of the deteriorating visibility. However, the advantage which would have accrued from the deviation around Los Angeles was largely offset by the subsequent climb to higher altitude. This placed EuNo458 in the airspace segment normally used by eastbound traffic climbing to the high-altitude route structure.
Two other decisions by the pilot of BuNo458 also had a the collision. First he did not significant effect on attempt to request radar traffic advisories. This would have alerted the appropriate controller that a - 18 - nontransponder target was in the area and undoubtedly would have resulted in an attempt to establish radar identification. Even if radar contact bad not been accomplished at that time, the general location would have been established, and traffic advisories could have been issued accordingly. In spite of the fact that crews sometimes do not sight the traffic even though the advisories are issued in specific terms (clock code and distance) the issuance of a general warning (geographic location and direction of flight) would have served to narrow the field of search, thereby increasing the probability of detection.
Secondly, he requested the RIO to conduct a radar mapping exercise at a time when he was traversing an area of dense traffic. Although it may be argued that outside visibility from the rear cockpit is relatively poor, all possible assistance in maintaining a lookout should have been used. If any radar exercise were to have been conducted, it should have been in the search mode. Even in the degraded condition of the radar, this would have been preferable.
2. Operation of RW706 An analysis of the final 0.6 second of flight recorder operation shows that at 8 minutes 54.6 seconds, the vertical acceleration transducer sensed a ~ositive g force, moving from +0.74g and culminating in a +5.08g reading at 8 minutes 55.2 seconds. At the instant the stylus recorded this, it moved instantaneously to a -1.Sg reading. The return to, and overtravel beyond, a +lg (normal) position (-1.8q), with no measurable elapsed time, strongly suggested rapid response or normalizing of the spring-restrained seismic mass in the electromechanical transducer after high excitation in the positive direction. These g recordings were made possible by the recording rate of 10 per second for this parameter, whereas the other three parameter rates are one per second.
The Safety Eoard believes that these excursions on the acceleration trace resulted from shock loading at impact and not from any attempted evasive maneuver by RW706.
- 19 - Since the crew of RW706 took no evasive action prior to the col1ision. this indicates that either they did not see BuNo458 or saw it too late to take appropriate action.
There are several factors which individually or co1lectively could have reduced the ability of the DC-9 crew to see and avoid the F-4B. The crew probably engaged the autopilot to maintain climb schedule and, under radar control. probably expected traffic advisories of converging targets from the controller. Further reduction in outside vigilance might have resulted from such normal cockpit functions as determining or changing various radio frequencies. adjusting settings or controls of the flight director or thrust levers. However. the probable reasons why the RW706 crew did not see BuNo458 were: (1) both aircraft had a nearly constant relative bearing to each other; (2) the high closure speeds; (3) the lack of conspicuity of Eullo458; and (4) the lack of recurrent training in etficient lookout doctrine and scanning techniques.
(d) Human Factors in Target Detection and Assessment The Board's cockpit visibility study (Attachments 3 and 4) indicates that at least 40 seconds prior to impact, BuNo458 was less than 450 to the left of the DC-9 captain's and first officer's normal sight line. Ap~roximately 35 seconds prior to col1ision. RW706 completed a left turn and was then climbing on a constant heading. Although the target size of BuNo458 was small at this time (0.017-inch).
the smoke trail from engine exhaust would have at least tripled the effective target size. The visual angle subtended by such a target would be approximately 10.8 minutes of arc. The empirically derived threshold for detection is nominally 4 minutes of arc. At approximately 15 seconds before the collision, just prior to the onset of the "blossoming effect" which occurred as the intruder target size increased dramatically, the size of BuNo458 and a smoke trail twice its length would have grown to ap~roximately 0.117-inch. In the next 10 seconds, the target size would triple, in the last 5 seconds it would expand to fill the entire visual field. These figures are predicated on the constant foreshortened length of BuNo458 which wou1d result from the relative positions of the ttio - 20 - aircraft, and no attempt was made to adjust for the target size during the evasive action taken by EuNo458.
The cockpit visibility study (Attachments 3 and 4) also indicates that RW706 would have been approximately 390 to the right of the normal sight line of the pilot of BuNo458 and approximately 370 for the RIO, for the last 40 seconds prior to collision. The target size of RW706, allowing for the foreshortening due to angular displacement, 35 seconds 'prior to impact was approximately 0.037-inch. At this point the target would subtend approximately 1 minutes of arc, which is well within the detectable threshold mentioned earliera No addition to target size was made for engine exhaust because one engine was modified, which reduced the visible smoke emission. This would have presented less than optimum density for detection. During the final 15-second period prior to collision, the target size of RW706 blossomed rapidly. tripling in size between 15 seconds and 5 seconds and then expanding to fill the entire visual field in the final 5 seconds.
Although the F-4B and DC-9 target images were t~eoretically of sufficient size to permit detection at 35 seconds prior to collision a number of factors could have contributed in this case to reduce the likelihood of detection at that timea In the analysis of any midair collision, laboratory data on human response and capabilities ~/ must be adjusted to real-world conditions~ The extent to which these data vary depends on the effect of many factors: ieeee windshield refractance, surface irregularities and cleanlinessg size and location of windshield frames, the background against which a target is viewed, atmospheric light scatter, and viewer training, ability, and preoccupation. All may be involved to varying degrees at the time a target is within a perceptible thresholdc The extent to which these factors affected detection of RW706 or BuNo458 cannot be determined preciselyc However, the various studies comparing laboratory data to real-world situations show dramatic reduction in the probability of visual detection due to the .factors listed above~ A nonstructu.red or ill-defined homogeneous background presents a less-than-desirable field when the search for a target is conductedo The lack of defined background texture, coupled ~ith a constant background hue, can severely limit not only the detectability of a target but - 21 - also the ability to perceive target motion, once the target is detected. While the effects of atmospheric light scatter, and the reported haze layer at 9,000 feet cannot be quantitatively determined in this accident, it is reasonable to surmise that RW706 presented less than optimum conspicuity when viewed against the haze layer. Moreover, its motion relative to the background haze would be difficult to detect. Had either aircraft displayed high intensity strobe lights, the increased conspicuity probably would have enhanced early detection of each aircraft.
Another factor which can affect the detectability of airborne targets is the myopic nature of the human eye when an air-to-air search is being conducted. The condition results from the tendency of the eye to focus at ap~roximately 20 feet during a visual search into an essentially empty visual field. Although this condition is more prevalent at extremely high altitudes where the horizon becomes ill-defined and high ambient lighting becomes a factor it is also possible that a myopic condition could exist at markedly lower altitudes when a pilot is searching against an ill-defined homogeneous fielda The possibility therefore exists that the crews of RW706 and BuNo458 could have been subject to some degree of myopic vision with a resultant reduction in their ability to detect a small target ..
Finally, the effectiveness of crew scanning is dependent on training and the time sharing of activities inside and outside of the cockpit. Based on a fixed-eye reference point, neither target was masked by intervening cockpit structure for any significant period of time; however, each target was in the peripheral visual field ..1.Q/ of all crewmembers.. The lack of relative motion of either target in the peripheral vision of any crewmember could have made early detection of the other aircraft highly unlikely.
Similarly, the small size and lack of relative movement of either target, even though detected at 35 seconds prior to collision, would undoubtedly have precluded accurate assessment of the vertical and horizontal separation or rate of change of target size. Thus even if the tC-9 and F-4 crews detected the other aircraft, the cues for accurate assessment of the collision geometry could have been marginally adequate.
It may be postulated that as the closure distance decreased from 20 to 10 seconds prior to the collision the - 22 - target would become better defined and Rw706•s climb attitude could be more accurately discerned by the pilot of BuNo458. Thus a sighting during the period between 20 and 10 seconds prior to collision might not have been interpreted as an imminent collision threat because of the smallness of the target size. However, target size notwithstanding, the fixed bearing of RW706 and its location near the horizon would have suggested that a collision threat existed and that he should maneuver to assure a comfortable separation. Moreover, the F-4B pilot's military flying experience, including tactical intercept training should have increased the likelihood of the initiation of a right turn or other maneuver which would have increased the miss-distance. The lack of any such maneuver indicates that he did not sight the DC-9 in sufficient time to have executed an appropriate maneuver to avoid the collision.
In light of the above discussion of the likelihood of early detection, the Safety Board concludes that although detection of RW706 by the pilot of EuNo458 might have occurred as much as 35 seconds prior to collision, it is more likely that it occurred at some time markedly less than 20 seconds prior to the collision.
The possibility of an early detection of BuNo458 by the crew of RW706 was considered. However, with BuNo458 located near the horizon and on a constant or nearly constant bearing, early detection probably would have ~rompted the crew of RW706 to monitor the progress of BuNo458 thereafter and seriously to consider altering their climb schedule or heading to ensure safe passing separation. Assuming continued assessment by the RW706 crew, as the range decreased, the likelihood of their making a ~recautionary alteration in flightpath would seem to increase. Therefore, in the absence of any such deviation in flightpath, the Safety Board concludes that it is most likely that the crew of RW706 never saw BuNo458, or saw it moments prior to the collision and had no time to initiate an evasive maneuver~ In order to determine a likely time for detection of RW706 by the pilot of BuNo458, it was necessary to consider the RIO•s warning coincident with the rolling maneuver as a starting point, together with aircraft response times and laboratory data suggestive of pilot response times in collision.situations. The data suggest that it would take 0.24 second to accommodate to foveal vision, cnce a target was detected. Neural processes would take an additional 0.3 - 23 - second. The data further suggest as much as 3 seconds could have elapsed during recognition and assessment of the various cues and determination that a potential threat existed. Approximately 2 seconds could have then elapsed while deciding whether an evasive maneuver was necessary and if so, the type of maneuver to initiate. Another 0.5 second could have elapsed for human motor response. Aircraft performance data indicate approximately 3 seconds could have been required for aircraft response, depending en the rate and type of control input. Based on the RIO's testimony and analysis of other events, the pilot's partici~ation in the radar mapping exercise was completed approximately 20 seconds prior to the collision. However, this remaining 20 seconds was most likely not entirely spent in constant visual search of surrounding airspace. such intracockpit duties as monitoring the attitude indicator to maintain flightpath attitude, airspeed, and status of aircraft subsystems would have occupied some finite amount of this time. Thus, the time available for detecting any outside target could have been significantly less than 20 seconds.
It is postulated that 10 seconds could have been spent performing a noncontinuous visual search of the surrounding airspace, while the remaining 10 seconds were shared with scanning cockpit dis~lays. Because the DC-9 target was very small, stationary, and located in bis peripheral vision, it is most likely that the pilot did not see the DC-9 until just moments before the collision. The completely unexpected appearance of the DC-9, together with its dramatic growth in size during the 10 seconds prior to collision rendered proper assessment of the situation extremely difficult if not impossible. The Safety Board concludes therefore, that it is likely that the pilot of BuNo458 detected RW706 less than 10 seconds before the collision and that the evasive maneuver was initiated approximately 2 to 4 seconds before collision. Within the final remaining 2 to 4 seconds a left roll was made as an attempt to avoid a collision. A more appropriate maneuver consistent with previous training would have been a roll to the right to increase miss-distance. However, the Board cannot determine with certainty that even this type of maneuver would have assured safe passage of the F-4 ..
The Board further concludes that the visual cues for accurate assessment of the collision geometry by the pilot of BuNo458 probably were inadequate. Then, when target range had been reduced sufficiently to afford improved visual cues, the time remaining was so brief as to make
section 91.67 of the Federal Aviation Regulations (FAR)
- 24 - unduly difficult the accurate assessment 0£ the geometry and proper response.
(e) Consideration of See and Avoid Concept section 91.67 of the Federal Aviation Regulations (FAR) jj/ places the burden on both crews to see and avoid other aircrafto Assuming detection of the other aircraft, FAR 91~67(c) placed an additional responsibility on BuNo458 to respect the right of way of RW706 Nonetheless, as can be appreciated from the foregoing analysis of this collision, the likelihood of a pilot's either not seeing an intruder at all or seeing the intruder and misinterpreting visual cues and then attempting an evasive maneuver based on incomplete visual cues, is highly probable. The problem-solving process required of pilots in these situations is often highly complex, and in many cases the problem is impossible to solve in time to avoid a collision. This is demonstrated by the fact that the crew of BuNo458 had received recent training in lookout doctrine and scanning techniques but were unable to avoid the collision.
conversely, the crew of RW706 received no formal company training on lookout doctrine or scanning techniques, and no such training is required by either the company or the FAA.
Although Air west pilots are evaluated for "alertness", this evaluation appears to encompass conditions inside the cockpit as well as outside. There are no definitive criteria to determine how effectively a pilot maintains a proper lookout. It may be argued that previous military training in lookout doctrine and scanninq techniques, coupled with years of flying experience, would result in excellent time-sharing for responsibilities inside and outside. However, it is equally true that years of experience without constant review and improvement would result in establishment and reinforcement of im~roper habit patterns. overcoming such a behavioral pattern, which involves no conscious process, would require a concerted retraining program with periodic recurrent trainings The Board believes it significant that there is no indication that the crew of RW706 ever sa~ BuNo458 under these circumstancesG The Board, therefore, reiterates the position taken many times before that for certain - 25 - operational conditions, the "see and avoid" concept is simply inadequate and the development of collision avoidance systems must be vigorously pursued.
Whereas this accident resulted from high closure rates andQ consequently, small target size until shortly before the collision, the Board also recognizes the more common type of midair collision occurring between aircraft at relatively low closure rates. The Board believes that for this latter type of collision, the detectability and assessment of the collision threat from an intruding aircraft can be enhanced by proper pilot techniques and a more thorough understanding of visual phenomena. The Safety Eoard•s publications _!1/ related to midair collisions between aircraft in visual meteorological conditions have stressed the need for increased pilot vigilancec Recommendations have been sent to the FAA, the air carriers, commercial operators, pilot associations, and the many aviation-oriented interest groups to increase the awareness of pilots to the midair-collision threat. It is therefore gratifying to see that many ox the professional ~ublications and meetings of these organizations are focusing on the many facets of this problem.
Similarly, a terminal control area has been implemented in the Los Angeles area, since the accidente This action is a positive step toward reducing the threat of midair collisions, but the Board believes the conce~t would not prevent the recurrence of this accidenta Esta~lishment of climb and descent corridors, as previously recommended by the Board would tend to eliminate this type of accident~ 2 .. 2 Conclusions (aa Findings 1. Both aircraft were airworthyo 2G All f lightcrew members were qualif iedo 3o RW706 was operatinq in accordance with an IFR flight plan under radar control of the Los Angeles ARTCC ..
4o BuNo458 was operating in accordance with a VFR flight plan and was not under control - 26 - of the ATC system.
5. The air traffic controllers were qualified for their assigned duties.
6. BuNo458 was not detected on radar because of an inoperative transponder, the aircraft radar cross-section, and a low level temperature inversion in the area.
1. There was no restriction to in flight visibility in the area of the accident.
8. The pilot of BuNo458 exercised ~ocr judgment in performing an aileron roll, but the roll did not contribute to the accident.
9. The pilot of BuNo458 attempted to· eject from ~he aircraft, but he was unable tc do so because the forward canopy did not jettiscn.
10. If BuNo458 had requested radar traffic advisories, the controller could have advised RW706 of the presence of BuNo458 and the probatility of avoiding the collision would have increased significantly.
11. USMC flightcrews receive training in lookout doctrine and scanning technique.
12. No formal training or evaluation of crew scanning technique and lookout doctrine is accomplished by Air West.
13. Both aircraft were theoretically cf sufficient size to permit detection by each other at 35 seconds prior to collision. However, detection and assessment were probably compromised by target size due to high closure rate, target contrast and location in the peri~heral visual field, and other visual limitations.
14e At 35 seconds before impact, both aircraft were on an essentially constant relative bearing and would have been difficult to detect because each target would be near the minimum detectable size and would remain relatively - 27 - stationary.
15. In view of the absence of evasive action on the part of RW706 (i.e., no alteration of heading, climb profile or airs~eed) it is logical to conclude that the crew did not sight BuNo458 in time to initiate such evasive action.
16. The pilot of the F-4B probably first observed the target of the DC-9 at about 8 to 10 seconds prior to collision, devoted the first portion of this brief period to assessing such cues as relative bearing, speed, and climt angle, and initiated a reflex evasive maneuver approxi- mately 2 to 4 seconds prior to the collision.
(b) Probable Cause The National Transportation Safety Board determines that the probable cause of this accident was the failure of both crews to see and avoid each other but recognizes that they had only marginal capability to detect, assess, and avoid the collision. Other causal factors include a very high closure rate, comingling of IFR and VFR traffic in an area where the limitation of the ATC system precludes effective separation of such traffic, and failure of the crew of BuNo458 to request radar advisory service, particularly considering the fact that they had an inoperable transponder.
3. RECO~ENDATIONS As a result 0£ this accident the National Transportation Safety Board recommends that the Administrator of the Federal Aviation Administration: 1o Install video tape at all FAA ATC radar displays, both term~nal and en route for use as an investigation tool. (A-72-200) 2 .. Install an open "area" microphone at each terminal and center sector position to record all conversation at the control positions. (A-72-201 3. Establish climb and descent corridors extending from the top of the TCA's to the base of APC, to remain in - 28 - effect until the base of APC has been lowered to the top of the TcAas. (A-72-202) 4o Establish more definitive procedures for the guidance of controller personnel in handling Code 7700 aircraft.. (A-72-203) So Review radar performance monitoring procedures to assure that all radar facilities are capable of receiving Code 7700 transponder returnsa (A-72-20q) Board also The National Transportation Safety the following recommended (A-71-52) that the FAA take action: coordinate with the Department of Defense, and, in areas where a large intermix of civil and military traffic exists, develop a program to insure that appropriate graphical depictions of airspace utilization and typical flow patterns are prominently displayed at all airports and operational bases for the benefit of all airspace users.
The Federal Aviation Administration responded, in a letter dated November 10, 1971, as follows: "This is in response to your safety recommendation, A-71-52, issued 9 November 1971.
"Recommendation number 4 of our Near Midair collision Report of July 1969 is similar to your recommendation ..
"As a result of that recommendation we: 1o Developed a new Fart 4 of the Airman's Information Manual in January 1970 (Graphic Notices and Supplemental Information). As graphics are made availatle, they are included in the semiannual Part 4 or are carried in the every 28-day Part 3, witil they can be transferred to Part 4 ..
2. Developed various types of graphic displays of normal IFR and VFR routes.
- 29 - 3. Developed graphics for the 22 large terminal hubsa Fifteen are published in the AIM, and the remaining seven are in various stages of processing.
4. Developed graphics for other than large huts including Air Force Bases. Eight are published in the AIM and three are in the final stages of development. Nine of the eleven display Air Force Base activities.
"In addition to the above, we have developed a VFR Terminal Area Chart (copy enclosed) which we will be testing and evaluating. This chart of Chicago Area depicts the Terminal Control Area, VFR and IFR routes and military operations at NAS Glenview. The chart is designed for use by pilots and for display at all airports and operational bases.
"As you can see, we have expanded upon the original recommen- dation in our Near Midair Collision Report. As a continua- tion of this expansion, we will coordinate furtber with the Department of Defense to incorporate ot~er military bases into the program.
"After review of the material we have outlined above, we would appreciate any further comments' you may have con- cerning this matter ..
On February 8, 1972, the Safety Board further recom- mended (A-72-12 & 13) that the FAA: 1. Develop VFR Terminal Area Charts, similar to that prototype portraying the Chicago TCA, for all other TCA s and, if feasible, for other large air traffic hubs a 2. Initiate a program to publicize the existence of, and the location of, these graphics for prospective users. The program should incorporate, in part, Examograms and, in airmen examinations, questions referring to these graphics.
The Federal Aviation Administration concurred with these recommendations in a letter, dated February 15, 1972.
Other recommendations (A-71-48 thru 51) were sent to the Department of Defense on November 2, 1971, suggesting the following actions: - 30 - 1. Review the feasibility of restricting all types of low-level training, which requires airspeeds in excess of the FAR limitations, to designated restricted areas and low-level navigation routes.
2. Rephrase the wording contained in your altitude/ airspeed limitations, and delineate explicitly those instances wherein airspeeds in excess of the 10,000 feet/250 RIAS limitations are authorized.
The Board believes that the exceptions should be limited to the following: "a. Climbs and descents to traffic patterns, authorized and/or designated training areas and low-level navigation routes.
"b. Those instances where safety of either crew or aircraft require operations in excess of the limitation.
3. Explore the feasibility of using the air intercept radar on all military aircraft to provide collision avoidance assistance as an additional aid to the see and be seen" concept; and should this prove feasible, institute and establish procedures to use the radar for this pirpose on all flights where its use is not required for more urgent military mission requirements.
4. Institute a program to provide more publicity to the existence, function, and use of the FAA Radar Advisory service in those instances where VFR flight is required through high-density traffic areas. consideration should be given to making the request for such service a mandatory ~rocedure.
The Department of Defense responded, in a letter dated December 2, 1971, as follows: "This letter is in res~onse to the National Transportation Safety Board safety recommendations A-71- 48 thru 51, which you forwarded to Secretary Laird on 9 November 1971.
"These recommendations have been referred to the military services for their consideration. I am advised that during their initial review the recommendations - 31 - were considered sound and would be implemented to the extent feasible. The details of such action are being staffed. The results of this staffing will be the promulgation of specific instructions and guidance to their operating commands.
"Thank you for your helpful recommendations which are so important to our mutual interest in achieving the greatest degree of air safety."
The safety Board previously made recommendations on the problem of midair collisions in the Board's special accident
prevention study "Midair Collisions in u. s. Civil Aviation
1968" which was released in July 1969, and the "Report of Proceedings of the National Transportation Safety Board into the Midair Collision Problem - November 4 through 10, 1969" which was released February 22, 1971.
BY THE NATIONAL TRANSPORTATION SAFETY BOARD: /s/ JOHN H. REED Chairman /s/ FRANCIS McADAMS H· Member /s/ ISABEL A. BURGESS Member WILLIAM R. HALEY /s/ Member LOUIS M. THAYER, Member, was absent, not voting.
September 22, 1972 - 32 - FOOTNOTES j/ All times herein are Pacific daylight, based on the 24- hour clock.
2/ Airspace within which all traffic is under positive control, and all aircraft must operate in accordance with Instrument Flight Rules (IFR). At the time of the accident, the positive control began at Flight Level 240.
3/ A collocated very bigh frequency omnirange and ultrahigh frequency tactical air navigational aid. The DME feature gives a slant range measurement to the facility.
4/ A controller qualified in the type of work being done, i.e •• radar, data, tower, etc., but who is not checked out in the specific position of a facility, i.e •• R-18, D-18, etc.
5/ MTI is a feature of the display which tends to eliminate returns from stationary targets. It is infinitely adjustable within the range capability of the radarscope, and has six preselected levels of signal attenuation available.
6/ PRF was designed to virtually eliminate any blind speed effect which could occur when targets are traveling tangent to the antenna, within the range of the MTI selection. Such targets would otherwise not appear on the radarscope due to apparent lack of motion.
7/ code 7700 is a universally used emergency code for transponders.
j!/ Guard channel is the international emergency frequency.
It is 121.5 MHz for very high frequency (VHF) communi- cations.
~/ A study of Requirements for a pilot Warning Instrument for Visual Airborne Collision Avoidance, Sperry Gyroscope Company, Great Neck, Long Island, December 1963; and Vision In Military Aviation, J.W. Wulfeck, et al., WADC Technical Report 58-399, November 1958, Wright Air Developm~nt Command, Wright-Patterson AFB, Ohio.
- 33 - .lQ/ Foveal vision takes place within 20° of the center portion (fovea) of the eye. Peripheral vision occurs outside this 200 cone of foveal vision • ..11/ FAR 91.67 states in part: "When weather conditions permit, regardl.ess of whether an operation is conducted under Instrument Flight Rules or Visual Flight Rules, vigilance shall be maintained by each person operating an aircraft so as to see and avoid other aircraft in compliance with this section. When a rule of this section gives another aircraft the right of way, he shall give way to that aircraft and may not pass over, under, or ahead of it, unless well clear." FAR 91.67(c) provides that: When aircraft of the same category are converging at approximately the same altitude ••• the aircraft to the other's right has the right of way ••••
12/ Midair Collisions in u.s. Civil Aviation-196~; Aircraft
Accident Report NTSB-AAR-69-2; Aircraft Accident Report NTSE-AAR-69-4; Aircraft Accident Report NTSE-AAR-70-15; and Report of Proceedings into the Midair Collision Problem NTSE-AAS-70-2.
APPENDIX A
- 34 - APPENDIX A INVESTIGATION AND HEARING 1. Investigation The Board received notification of the accident at ap~roximately 1830 on June 6, 1971, from the Federal Aviation Administration. An investigating team was immediately dispatched to the scene of the accident.
working groups were established for Operations Air Traffic Control, Witnesses, Weather, Buman Factors, Structures, Powerplants, Systems, and Flight Recorder. In addition the maintenance records for each aircraft were reviewed. The Federal Aviation Administration, Department of Navy, Hughes Air West, McDonnell-Douglas corporation, and Air Line Pilots Association all participated in the investigation as interested parties. The on-scene investigation was completed on June 19, 1971.
2. Hearing A public hearing was held at Pasadena, California, on July 27 to July 30, 1971. Parties to the Hearing included the Federal Aviaticn Administration, Department of Navy, Hughes Air west, McDonnell-Douglas corporation, Air Line Filots Association and Aircraft owners and Pilots Association.
3. Reports A preliminary factual report of the investigation was released by the Board on July 26, 1971. A summary of the testimony was issued on August 20, 1971.
APPENDIX B
- 35 - APPENDIX B Crew Information Captain Theodore Nicolay, aged SO, held airline transport pilot certificate No. 474317, with ratings in airplane multiengine land, F-27, DC-3, and DC-9. Be had accumulated 15,490 total flying hours, including 2,562 hours in the DC-9. Be completed his last ~rof iciency check on March 8, 1971, and his FAA first-class medical certificate was issued on January 25, 1971, with no limitations.
First Officer Price Bruner, aged 49, held airline transport pilot certificate No. 611777, with ratings in airplane multiengine land, F-27/227, DC-3, DC-9 and commercial privileges in airplane single-engine land. Be had accumulated 17,128 total flying hours, including 272 hours in the DC-9. He completed his last ~roficiency check on December 7, 1970, and his FAA first-class medical certificate was issued on December 18, 1970, with no limitations ..
The f lightcrew members had been on duty ap~roximately 6 hours, including about 1 hour 50 minutes f ligbt time, when the accident occurred. Their rest period prior to reporting for duty was 18 hours 13 minutes.
Hostess Joan R. Puylaar, aged 34, was hired on August 22, 1959.
Hostess Particia Shelton, aged 28, was hired on October 18, 1963.
Hostess Helena Koskimies, aged 30, was hired on October 16, 1967.
All cabin crewmembers completed their prescribed emerqency recurrent training in September 1970.
First Lieutenant James R. Phillips, aged 27, held commercial pilot certificate No. 1619834, with ratings in airplane single- and multiengine land. He also held a valid flight instructor certificate. Be had accumulated 440 total - 36 - military flying hours, including 170 hours in the F-4B. He also had accumulated approximately 400 hours in civil aircraft. He completed his last Naval Air Training and Operating Procedures Standardizations Programs (NATOPS) flight check on December 8, 1970, possessed a Standard Instrument card, and passed bis annual medical examination on June 23, 1970, with no limitations.
First Lieutenant Christopher E. SChiess, aged 24, joined
the u. s. Marine corps on November 7, 1969, and completed
his RIO training January 3, 1971. He had accumulated 195 total flying hours, including 89 hours in the F-4B. He was current on all prescribed training and passed his annual medical examination on July 13, 1970.
Both crewmembers had been on duty approximately 1 hours 4 minutes, including 1 hour 53 minutes flying time at the time of the accident. Their rest period prior to reporting for this flight was 19 hours 10 minutes.
APPENDIX C
- 37 -
APPENDIX C
Aircraft Information
N9345, a McDonnell-Douglas DC-9-31, serial No. 47441,
was owned by the C.I.T. Corporation, 650 Madison Avenue, New
York, New York, 10020, and operated by Hughes Air West, San
Francisco International Airport, San Francisco, California,
94128. It had been flown a total of 5,542 hours at the time
of the accident. A review of the records indicated that all
applicable Airworthiness Directives either had teen complied
with or were scheduled for completion. Pratt & Whitney
JT8D-7 engines were installed as follows:
Position serial Number
Time Since overhaul
P65704-D
5265.69
P654152-B/D
2263.06
The aircraft weighed 86,518 pounds at engine start and
the center of gravity was 21 percent MAC. Both are within
the allowable limits.
Bureau Number 151458, a McDonnell-Douglas F-4B, was
received on April 15, 1964, and had been operated by various
squadrons of the u. s. Marine Corps. At the time of the
accident, it was assigned to VMFA-323, and had teen flown a
total of approximately 2,03-0 hours. A review of the records
indicated that the aircra~ was maintained in accordance
with the appropriate regulations. General Electric J79-8
engines were installed as follows:
Position serial Number
Time Since overhaul
401437
842.5
421669
258.8
- 38 -
The aircraft weighed approximately 43,310 pounds at
engine start. Both the takeoff weight and center of gravity
were within prescribed limits.
ANGELES LOS .~ 16f.O .-,ppffoACH OR f
1! ~? ~~
IH -.NGELES ON
-·-f
LOS T TROL CO ICONT /y -1 ( ('
BOARD
N9345 151458
SAFETY
TAS DC-9-31. BUNO D.C.
420K 1971 CALIFORNIA F-411, 6, COWSIJN DOUGIAS IJNE CORPS DUARTE,
ATTACHMENT
MmAIR WASHINGTON, WEST NEAR AB!
TRANSPORTATION MARINE
U.S.
IWGHES
NATIONAL
RATES
CLOSURE
AND
PATH---- NORTH FLIGHT BEARINGS, TRUE DC-9 RANGES,
COMPUTED
MONYI WEST TAS IMG F-48 TEST! TRUE 145' 42DK IMNEUVER 0 I OF IR HORIZON 14'N THE ) EVASIVE WAS TO ABOVE
TAS ~
IMG 20' DC-9-31 RECORDER PRIOR 39° ~------------------4.JSNM--------------------J 400K POSITION AN.D.
FLT.
I SUN
BOARD
EYE N9345 DESIGN 151458 :-~·.71
SAFETY
ATTACHMENT 3 DC-9-31, BUNO DESIGN D.C.
OFFICER'S CALIFORNIA F-48, 6, COLLISION DOUGLAS FIRST JUNE CORPS CAPTAIN'S DUARTE, MIDAIR WASHINGTON WEST FROM POINT
TRANSPORTATION
NEAR AIR FROM MARINE U.S.
POINT HUGHES VISIBILITY
NATIONAL
REFERENCE VISIBILITY DC-9 EYE DC-9 REFERENCE WINDOW I
I
SIDE CO-PILOT'S I SS
I
I SO
I
I 4S
I
I
I
I 3S
I
RIGHT I RIGHT
30 I
· I 2S
I
I
I
I lS
I
DEGREES I DEGREES
I
- -
I
I
I
osmgw~~~~~~"~
EB
I ZERO REFERENCE
I
WINDSHIELD WINDSHIELD VISIBILITY I
I VISIBILITY
I
I
CAPTAIN'S FRONT CO-PILOT'S FRONT I
I
LATERAL I LATERAL
I
PATH I
I
LEFT LEFT I
I
I FLIGHT
I
I
I
I I I ~"~~~~~~wgmso CALCULATED ~"~~~~~~wgms .....
F-4 WINDOW SIDE ~ WINDOW SIDE CAPTAIN'S CAPTAIN'S C C
I I
10 10 - -
I I
0 0
EB
EB
ZERO REFERENCE ZERO REFERENCE WINDSHIELD WINDSHIELD
VISIBILITY I VISIBILITY
I
I
I
CAPTAIN'S FRONT CO-PILOT'S FRONT
I I
LATERAL LATERAL
I
I
25 25 PATH
I
I
30 30 LEFT LEFT
I
I
FLIGHT
I
I
I
I
45 45
I
I
I
55 I
I
CALCULATED F-4 WINDOW SIDE CAPTAIN'S
~ g
D.. :::>
a.. :::> z :;: 0 Cl
5- s- 5- 0- 5- o- 15- 10- 15- 10- 20- 20- - -15- -20- -25- -30- - -10- -15- -25- -30- -10- -20-
LU >
~ ;; Vi ~ 6 ~
<..? Cl >- § > ~ ;:: ffi > "" <..? LU Cl :J
V> LU LU "" LU I- :::; V> --' u V> LU LU
N9345
BOARD
151458
BUNO
SAFETY DC-9-31, D.C.
POINT POINT
F-48, 1971
CALIFORNIA 6,
COLLISION
DOUGLAS
ATTACHMENT
REFERENCE· JUNE DUARTE,
REFERENCE CORPS
WASHINGTON, EYE
EYE MIDAIR
WEST
NEAR
TRANSPORTATION
AIR
MARINE
DESIGN DESIGN U.S.
RIO'S PILOT'S
NATIONAL
HUGHES
FROM FROM VISIBILITY VISIBILITY F-4B F-4B I I I I I I I I I I I I I I I I I I I I 100· I I I I I I COLLISION COLLISION I I TO TO I I I I 7o RIGHT I PRIOR I PRIOR RIGHT I I SEC SEC I I I 50 I I I I I I I I 30 I I I 20 I I I DEGREES I I DEGREES 10 - - I I I o I I I I VISIBILITY I 10 VISIBILITY I I I I I I LATERAL LATERAL I I I REFERENCE I I I I I ZERO I OBSCURED 50 I ONLY I I REPRESENT VISION I I I REPRESENT I VISION AREAS I 7o I LEFT COMPLETELY LEFT I I AREAS IS I I CLEAR BINOCULAR I I 90 I SHADED MONOCULAR I I VISIBILITY I I I I I I I I I I I I I I I I o: I I I I 50 50 I I I I 40 40 I I I I 30 30 I I I I I I DEGREES DEGREES I I 10 - - I I I I o 0 I I I I 10 VISIBILITY 10 VISIBILITY I I I I 20 20 I I LATERAL LATERAL I I 30 30 I REFERENCE I I I 40 40 I I ZERO I OBSCURED I 50 50 ONLY I I REPRESENT I VISION I I REPRESENT VISION I I AREAS 7o 70 LEFT LEFT COMPLETELY I I IS AREAS I I 00 80 CLEAR BINOCULAR I I SHADED MONOCULAR I II VISIBILITY I I 100 100 I I I I I I I I I I I I II I I
II
140: 140 I I s- --25 --30 5>- --15 --20 -0>- --5 --15 --20· --25 --30 25- -0'- --5 --10 30>-- 25>- 20t- lOr- 30>- 20>- 15f- 10- z--10 ~ 0 0 z ~ ~ ~·1si- ' >
> .... lS ;:: ffi
~ ~ "' "' "' "' ~ ~ :::; ;; v;
~ ::; ;; v; > 6
"' "' ~ ~