The Investigation Process Research Resource Site
A Pro Bono site with hundreds of resources for Investigation Investigators
Home Page Site Guidance FAQs Old News Site inputs Forums
. . . .. . . . . . . . last updated 4/28/07

INVESTIGATING
INVESTIGATIONS

to advance the
State-of-the-Art of
investigations, through
investigation process
research.



Research Resources:

Search site for::


Launched Aug 26 1996.

 

See also NTSB's report of this accident.

Investigations of the Loss of
Lockheed C130A Tanker 82
Pearblossom, California
13 August 1994

(Previously published in Wildfire 8:2, February 1999 and in the Journal of World Investigators Network, Spring 1999.)

Douglas R. Herlihy, Air Safety Investigator


Abstract

This paper briefly outlines the investigation and the re-investigation of the crash of a C-130 Hercules firefighting airplane that crashed in 1994 in the mountains of California while enroute to a fire call. The aircraft was destroyed and the crew of three were lost. A government investigation reported in 1995 that fuel had probably leaked in the center wing section and was ignited by electrical wire arcing. Three years after the mishap, in 1997, an investigative team climbed to the undisturbed wreckage site to re-examine the evidence. What they did not find, lead to reconstruction and re-examination of critical aspects of the event, and resulted in quite another conclusion. That independent investigation suggests another series of factors than the fuel-ignition-explosion scenario, and asserts that the evidence strongly supports an event in which inflight structural failure precluded the ignition of wing tank fuel. The compromise of major wing skin structure, set in motion a rapid chain of in-flight breakup events, in which the wing fuel ignition was simply most noticeable by ground and other airborne witnesses.


Firefighting air tankers - high risk - little oversight - and a history of losses

Since the 1950s, military bombers and cargo aircraft have been modified as air tankers, dropping slurry mixtures on forest and range fires, in most all states, including Alaska. The small industry began with “crop dusters” [1] and expanded into WWII single engine Grumman TBMs and TFs to be followed larger planes such as a B-17 and Neptune P2V patrol bombers, now being replaced by Lockheed P-3 Orions and other specialized aircraft. Smaller military aircraft, such as the Grumman S2 are being re-engined [2] for use into the next century. Many older commercial airliners are still used as well, including the DC-series 4, 6 and 7. In the late 1980s the U.S. Forest Service, through civilian contractors, released Lockheed C130 transports from storage which were then modified [3] with large internal water tanks.

The 1994 wild fire season in the United States was especially long and active. Early in the year, forest and range fires raged in the southwest, and as they were controlled, fires in Idaho, Montana and Oregon flared. In the air-tanker industry, accidents claimed a Lockheed P2V near Missoula, Montana [4] in July, followed by the loss of the C130A which is the subject of this paper. These planes, like other former military airplanes, modified as “borate bombers,” dominate the inventory of those used by companies who contract annually with the Forest Service. Specifically designed fire fighting aircraft have been limited in production. Accidents rates have been very high and while the rate per flight hour has not been established, over 70 firefighting flight crews have been lost in this small industry since the 1960s. [5]

Investigations into air tanker crashes have been cursory, and in review, a large number of crew-related causal factors seems apparent, suggesting events wherein task saturation and loss of situational awareness played major roles. Investigations as to the causes of these too often have suffered from a lack of government expertise, interest and resources. For many years, the government agencies that managed the contracts, dispatched the aircraft and detailed the training, also investigated the accidents. It is also apparent that these agencies’ actions or inactions contributed to the high accident rate, and recent federal court decisions have found that the agencies do, indeed, play a major role in the accident cause. [7] Many accident causes have been listed as “wing came off” or “aircraft collided with terrain,” which, of course, describes what happened, not why it happened. It is similar to listing a homicide cause as “gunshot wound.”


Background summary - the government investigation into the C-130 loss

On August 13, 1994, at around 13:10, a civilian Lockheed C-130A, on contract to the California Division of Forestry and the U.S. Forest Service, departed its base at Hemet Valley in Southern California on a firefighting mission. The crew of three, climbed to 8500 feet on a visual flight plan and took the tanker-modified aircraft [8] northwest toward the location of a wild fire on the slopes of the Tehachapi Mountains in southern California. About twenty minutes later, the right wing was seen to separate at the wing root in an orange fire ball and black smoke. The location of the break-up showed the aircraft had just entered into a region of ridgelines that rim the southern end of the Mojave Desert. Following a short radio squeal and a broadcast expletive believed to come from the aircraft, the airplane broke apart with a fire ball and impacted near the 7025 foot ridgeline. [9]

The 1956 transport was destroyed by impact and post crash fire. There were no survivors. The U.S. National Transportation Safety Board investigated and issued a probable cause and report, under a recent change to regulations, [10] allowing them to investigate “public-use,” [11] aircraft. Their findings indicated a fuel explosion from electrical ignition. Probative evidence in records or at the accident scene failed to support a fuel-air explosion scenario. Nevertheless, the majority of eyewitness, (some quite technically qualified as observers), [12] reported a flash and explosion that separated the wing.

The airplane was not equipped, nor was it required to be under regulations or contract, with a cockpit voice recorder (CVR) or flight data recorder (FDR). The “inflight explosion” scenario began to take on more importance, as more witnesses were located, further adding to the wing fuel explosion theory. And, according to the NTSB report, the supporting Lockheed representatives told of an earlier aircraft loss in a wing explosion which had its origin from a lightning strike. An area lightning/convective storm study by the Bureau of Land Management [13] did not find a high probability of lightning strike in this accident. [14] Furthermore, and quite significantly, witnesses reported the “smoke rising straight up” rendering a potential surface wind/weather factor moot. Weather was deemed not to be a factor. [15] Since the airplane had been modified with bladder tanks in the normally empty center wing sections, and pieces of material from these bladders were found at the scene, the hypothesis of cause seemed to focus early on wiring within that center wing box structure, that could have ignited fumes from such a bladder or fuel lines, had they leaked. The NTSB issued its report [16] finding probable cause to be a “fire/explosion” of the fuel system.


Factual findings of the Safety Board report

Wreckage distribution and on-scene investigation

Following the crash, the NTSB investigator, accompanied by an aviation safety officer from the US Forest Service, and a union representative from the Air-Tanker Pilots’ Association were landed on the ridgeline by helicopter. [17] They descended the sand and gravel slope, examined and photographed the scene. The NTSB investigator remained with the two overnight due to the onset of darkness, and he departed by helicopter the next morning. Other party members remained at the site during the rest of the second day, joined by manufacturers’ representatives. At the NTSB’s request, the Air National Guard removed the four (4) engines, [18] pieces of top center wing skin, and propellers number 3 and 4 from the right wing area [19] by long-line cable during the following week. They attempted, but failed, to lift the cockpit.

The initial on-scene investigation diagrammed the wreckage pattern in three areas stretching southeast to northwest parallel to the main ridge, the first being a location without fire damage containing six pieces of top center wing section, then at a distance of 1700 feet more, the burned right wing and engines, followed 300 feet later by the main wreckage of the fuselage, tail assembly, left wing and engines and the cockpit. The total wreckage pattern covered a distance of about 2000 feet, characterized by steep terrain and small mountain timber and brush with many burned-over areas. [20]


Site one, center-wing top skin

Ultimately, these six wing pieces were the most important parts of the aircraft to be located. The largest piece of the skin, [21] labeled as piece number one, was approximately 17 by 4 feet in size, mated with the other five smaller pieces. The number four piece still had the decal indicating that it had come from the area of the “AUX FUEL” filler cap on the left end of the top wing box. [22] While oily and dirty deposits were found on the wing pieces, the NTSB laboratory found no evidence of heat damage to the underside of these pieces. The laboratory report also noted that the surface pieces, numbers one and two, had “pulled away from the ribs.” Furthermore, the report stated that, “...fractured rib clips at the rib locations and most of the screws that attached the upper rib cap were still attached. The failure of the rib clips was consistent with a vertical or near vertical tensile failure.”


Site two, right wing

About 1700 feet northwest on a steep down slope, the remains of the right wing, engines and propellers were located. These engines and propellers were airlifted during the week following the accident by use of long-line cable under an Army Air National Guard helicopter. These right wing engines and propellers were examined by manufacturers’ technicians under the approval of the NTSB and were found to be unremarkable with few significant pre-existing anomalies. The engines all showed some indication of power and rotation at impact. The right inboard engine however did show two key differences. [23] The right inboard propeller, while mostly destroyed by post crash heat, was found to be in reverse. (It was noted that this engine was identified in the NTSB Factual Report as the right “outboard,” however, the Powerplant Group investigation correctly identified it as the number three or “right inboard” engine (closest to the wing root and point of wing separation). In addition to the propeller reversal, the “hot” [24] sections of engine 3 and 4 showed “heavy coking” or sooting of petroleum product. While not cited as significant findings in the NTSB report, these anomalies later have taken on more importance in re-examination.


Site three, fuselage, tail, cockpit and left wing

The fuselage, tail section, cockpit and left wing (with engines and propellers) impacted on a 45-degree slope approximately 300 feet beyond the right wing and 2000 feet along the track from the first location of the center wing section top skin. Approximate angles of impact could be measured from the left (numbers one and two) propellers to be 90 degrees to the slope. [25] According to the NTSB report the cockpit wreckage was attempted to be lifted, by helicopter long line absent the NTSB investigator (italics added) , but broke apart and was abandoned at the site. [26]


Witness summaries

Witnesses were interviewed by sheriff’s deputies, Forest Service employees and the NTSB investigator. The NTSB report summaries by reference to written or interview reports, nineteen (19) reports. Of the nineteen, fourteen reported the aircraft to have “exploded” with comments on various fireballs and smoke clouds. Of those, eight reported the wing separation and debris falling. Witnesses also reported that the smoke cloud rose from the site following the crash in an “undisturbed, vertical column.” [27] Half of the witnesses described the breakup sequence which was consistent with the impact angles of the number one and two engines (90 degree impact to the slope).


Weather not a factor

The convective storm summary and witness statements regarding the rising smoke columns seems to have supported the absence of weather as a factor in the NTSB accident brief. The likelihood of a lightning strike was discounted.


Maintenance record review

The NTSB assigned the investigation of the operator’s maintenance records for the accident aircraft to an FAA inspector from a Flight Standards District Office. Following his record inspection, the FAA inspector was flown to the site by helicopter on 18 August 1994 and photographed the site. His report indicates that the former military C130A aircraft, as operating in the restricted category and had no applicable Airworthiness Directives (ADs) to the C130A or the L-282. [28] The FAA report [29] reviews discrepancies and maintenance completed from the day of the accident, back to May of 1994. No fuel leaks or other major discrepancies were reported. He noted that the aircraft had entered military service in 1956 [30], and had logged 19,547 hours, and the operator had logged 742 hours of that since receiving the aircraft from storage in 1988. The report indicated that the operator had complied with provisions of the FAA-Approved Aircraft Inspection Program (AAIP) prior to placing the aircraft in service that year for firefighting. [31]



Safety Board finds that the C130 fuel system exploded

Fuel leaks and electrical ignition cited as probable cause

Nothing remained of Tanker 82’s wing root area, centerwing box or dry bays, other than the six pieces of wing skin found 2000 feet prior to the main wreckage, pieces of foam insulation identified from the dry bay area, and pieces of charred rubber bladder auxiliary tank. Wiring, fuel plumbing and valves located in this center wing section were destroyed.

An exemplar aircraft, in storage in Tucson, Arizona was examined by the NTSB in 1995. That aircraft had been in storage for six years. After examination of the “dry bay” area of that aircraft, dried and flattened “o-ring” seals [32] were retrieved from disassembled valves. That aircraft was also equipped with rubber auxiliary fuel tanks in the dry bay compartments. The wiring connections to the fuel valves within the dry bay compartment were connected with “cannon plugs,” [33] with the exception of bonding (aircraft grounding) wires, which were uncovered connections.

The Safety Board’s “Brief of Accident” stated in part, “ The dry bay area of the right wing contains high pressure fuel lines, unshielded and exposed electrical wiring, and is in close proximity to the number 3 engine. The main fuel tank is outboard of the dry bay. No lightning was reported in the vicinity of the aircraft. C-130 aircraft have a history of fuel leaks in the dry bay. The source of leaks, flattened or pinched o-rings are on-condition replacement items. The aircraft was in long term storage in the desert for 2 years prior to acquisition by the operator for fire tanker duties. The U.S. Air Force emergency procedures warn of fuel leaks in this area and require inspections prior to each flight.

Summarizing the findings of cause from the Safety Board, the report cited that occurrence of a “ fire/explosion” was due to a (1) “fuel system leak; (2) electrical system, electric wiring - arcing; (3) powerplant - other; (4) fluid, fuel - exploded.” Adding an occurrence of “airframe/component/system failure/malfunction,” the Board found that (5) fuel system, tank - exploded; and (6) the wing had separated.


1997 Independent re-investigation

Re-examination of wing skin and stringers

In 1997 a re-examination of the accident began with the lay-out of the center wing top skin fragments on top of and exactly, over the same wing stations of an exemplar C130A. While the original NTSB laboratory report had found no evidence of heat on the underside of the skin of the box, it was still undetermined what the dynamics of the overload failure had been. The wing box itself, all 440 inches long is a closed vessel, resembling a large, elongated anchovy flat, with open framed ribs throughout. Most of its interior dry bay is reachable though access plates in the wing. The investigative team recognized that any explosive over-pressure generated at one end, would, like a tin can, be distributed to the other end. Notwithstanding evidence of a lack of heat from the underside of the skin and stringers (which were shed over the first debris site), the pieces were measured and photographed to confirm any sign of overpressure from the underside. There was none.


Search for other signs of explosion or overpressure

In April of that year, a search for overpressure continued by examining the wreckage at the crash site. The two investigators reached the site by climbing from the bottom of Pechner Canyon which was accessible from a trail ending approximately two miles below the site. Due to the difficulty for the public to access the site from the bottom of the drainage, the wreckage appeared to have remained undisturbed since the time of the original investigation in 1994 that brought personnel by helicopter. Wreckage distribution was consistent with the original diagrams with the exception that the slurry tanks were not at the location of the fuselage impact, but were found to be flattened and fractured in a ravine 350 feet before the impact of the cockpit and 150 feet right of centerline from site one to site three. At Site 2, the internal slurry tanks and both right main wheels were found with parts of their vertical tracks and screw jacks. C130 main wheels, mounts and doors are fuselage-mounted, and their impact location provided additional information as to the likely yaw and break-up sequence before impact.

Investigators examined all fuselage skin, panels, access plates, and flooring, as well as components and framework for signs of any overpressure, bulging, compression or decompression signatures. Over 200 hundred pieces were examined and photographed, including pumps, actuators, accessory motors, tubing, wiring and interior structure. No signs of explosion or overpressure failure were found.

The cockpit instrumentation and overhead panel was found unburned, but fragmented, in a position down slope from the original impact point, consistent with the reports that the cockpit broke up on the attempt to retrieve it intact by helicopter long line. Meaningful information, at least in the field, in the condition the instruments were in, was not possible. The cockpit debris rested where it impacted the second time. Power quadrant, engine and flight controls remained, but due to its condition down slope, control position was no longer possible to determine.

Numbers one and two propellers and gear boxes remained impaled at right angles to the slope, their propeller blades fixed at some degree of traction (forward thrust, as opposed to reverse or feather), the teeth of the gear reduction units showing some rotational energy damage.

Cargo, from the aircraft was found, including chairs, camping equipment, tool chest, tools, tool boxes, chains, crew luggage and personal belongings.


The re-examination of the number three propeller - Analysis

Recalling that the NTSB Powerplant Group had found the number three (right inboard closest to the wing root failure) to have been in reverse, an examination was initiated into that anomaly. The NTSB report assigned no significance to that finding. Likewise, significance of the presence of “coking,” or unusual amounts of burnt oil/petroleum products in that engine were not addressed in the Safety Board’s report. Investigators now sought to examine this anomaly for a possible link to the event sequence.

Investigators utilized a classroom demonstration model of a propeller system, which had functioning levers, pulleys and cables to examine the operation of the propeller and throttle control system of the C130A. ( See photo insert). Additionally, a fully operational Hamilton Standard 54H60-90 series propeller and propeller system were examined and operated through its ranges and cable connections at a propeller rebuild facility in California. The cable movement “warning” information in the C130A manual [34] was consistent with movements of the actual cables to the propeller control; that is, that if a throttle (or propeller) cable is broken or moved without pilot input, it could cause the propeller to go to full power or reverse pitch. The accident propeller was found in reverse. Based on the reverse limiting horsepower (60 percent of full power), the aircraft’s airspeed, altitude, temperature (255 knots, 8000 feet, 39 degrees C), the number 3 propeller could deliver about 2250 horsepower in reverse. The “noise burst” radio transmission, believed to come from the aircraft, [35] is consistent with an initial catastrophic event at that moment and geographical location. A transmitted expletive was heard seven seconds later. [36] In that the engines govern at 100% RPM and the power requirements to achieve 255 knots would have been in the 900 degree range, [37] the propeller reversal from the throttle/propeller cable(s) break or stretch was likely part of the catastrophic wing separation. No accurate estimate of cable movement required for reversal was possible, though it is believed that as little as 1.5 inches (based on propeller control cam movement) could produce that result. Due to the variables of slack, tensioning device and the unknown dynamics of the breakup sequence, this measurement may be impossible to determine.


Tanker 82’s track and speed plot, and the wreckage location

The radar recording (NTAP file) was obtained from the FAA, as was the voice recording of the air-to-ground radio transmissions from and to the aircraft. The accident aircraft, though on a visual flight rules (FVR) flight plan (filed with firebase dispatch), was assigned a transponder code by the departure control facility as the aircraft climbed north toward the mountains. This transponder code could be identified in at least 12 “hits” in the data in the geographic area of the mishap. Plotting the aircraft northwesterly, the radar recording shows the aircraft moving at approximately 5,142 feet between 12-second “hits,” or at an average speed of 257 knots GS (nautical miles per hour, ground speed.) As Tanker 82 neared the site of its break-up, its speed slowed to 4962 feet between 12-second “hits” or an average speed of 248 knots.

Plotting the track of Tanker 82 from radar hits, without changing course, the wing debris (wreckage site one) could be plotted within 100 feet of (left) of track, 600 feet beyond the estimated position (based on 248 knots) of the radio transmitted voice call “expletive.” Wreckage site two (right wing) and the fuselage was found to be 2300 feet beyond that last radio transmission, and left of the extended track about 300 feet. These plots of radar recordings, overlaid on the quadrangle topographic chart, [38] computed locations of transmissions, places the wreckage pattern on the track of the aircraft, particularly associating the center wing debris with radio transmissions more than 2000 feet before wing separation.


Weather a factor

The NTSB report cited witness statements that “smoke was rising straight up” in undisturbed columns. An examination of weather data from the National Weather Service reporting stations revealed that while winds were indeed light and variable in the area (south Mojave desert mountains including, Palmdale, Edwards AFB, and San Bernadino to the south, the surface temperature at the time was 39 degrees Celsius (102 degrees Fahrenheit). Vertical winds with energy sufficient for at least moderate turbulence were calculated by climatological experts, and was consistent with variations in altitude of aircraft plotted on the NTAP. [39] Notwithstanding the allowances for accuracy, Tanker 82’s recorded altitudes (his intended altitude was transmitted to be 8500 feet), ranged on each successive radar “hit” to be varying between 8600 and 8300 feet, and down to 8200 feet on the last altitude record (second from last before radar loss). This is consistent with the meteorological analysis of at least moderate turbulence over this terrain at 102 degrees Fahrenheit.


Terrain and track

The radar plot and altitude record of 8200 feet (last record before loss) has the aircraft at 248 knots within 350 feet horizontal of topographic contour lines of 7800 feet [40] and crossing a series of southwest-northeast ridges of 7800 to 6800 feet at the time of the “radio noise burst” The site of the center wing skin and stringers, is located in the drainage beyond these ridges, as is the wreckage of the right wing and fuselage.


Airspeed limitations, operating weight and cautionary information

The accident aircraft was built in 1956 and accepted in 1957. Except for two years of desert storage to 1988, it had flown about 20,000 hours in 37 years of operation. Tanker 82 had departed Hemet Valley 20 minutes earlier with a full load of 3000 gallons of retardant mixture weighing approximately 30,000 pounds, [41] fuel and crew brought the weight conservatively to maximum gross weight of 120,000 pounds. The aircraft’s charted airspeed limit for that weight [42] was 198 knots indicated airspeed, but cautioned that the aircraft was turbulence limited to zero negative and +2.0 positive. These charts are, of course, based on a new aircraft, and refer to the “limiting load” (LL) [43].


Human performance and Crew Resource Management (CRM) considerations

The reality of aerial firefighting losses cannot overlook the fact that tanker captains are a community of strong individuals who see their firefighting role as similar to combat flying. Pilots are recruited by tanker companies, and the U.S. Forest Service contract supports, background qualifications of agriculture pilots (crop dusters) and requires the contractors to operate under 14 C.F.R. part 137 (agricultural spraying). Many captains have come through the ranks from single engine, single-piloted aircraft. Likewise, U.S. Forest Service contract overseers, with pilot qualifications, while many are multi-engine qualified, come from a background of light twin aircraft which are single-piloted. These include the Forest Service Beech Baron “lead planes.” [44] Cockpit cooperation and communications are known to be weak, with captains often communicating, flying and executing the drops with little task sharing or separation of duties. [45] Inasmuch as tanker captains hold the official “Initial Attack Cards” [46] and copilots traditionally have little input to the operations, this accident seems consistent with many other tanker accidents wherein operational limits have been exceeded, or cockpit situational awareness was lost. Without a cockpit voice recorder in Tanker 82, of course, we will never know the level of CRM in that cockpit, therefore nothing serves the evaluation to speculate on this crew’s interaction.

Beginning in 1995 and improving from season to season, the U.S. Forest Service and certain air tanker contractors have made considerable upgrades in training methodology. With more complex cockpits being operated, crews are preparing for seasons with simulator and CRM training, and more emphasis is being placed on cockpit communication. The urgency and hazards of air-drop missions will never be less of a factor, but risks can be (and are being) reduced with improvements in cockpit communication, cooperation and professionalism.


Making a case for reconsideration

NTSB evidence

The evidence presented by the Safety Board, that an “O-ring” that is always seated in a Wiggins fitting somehow “sprayed” Jet A jet fuel in a sufficient fuel/air mixture to detonate and blow the wing off, is not supported by universal maintenance knowledge. Neither is it supported by the aircraft’s maintenance history nor evidence found at the site. There are occasions when Wiggins fittings “weep or seep” fuel, but not spray. Lockheed and U.S. Air Force records do not support the Board’s reference to an aircraft history of a wing explosion (aside from the lightning strike accident in 1973).

* Anecdotally, any pilot or mechanic who has had to start a C-130 time and time again would attest that it is hard enough to get a fire started in the T-56 chamber with atomized fuel pressure, much less a dripping Wiggins fitting.

The Safety Board cited “bare wires” in the dry bay compartment of the centerwing section. There are no bare wires in the compartment, from wing station 220 right to wing station 220 left. All wires, except those bare bonding wires of ground potential , are protected by insulation and cannon plug, the barrels and locknuts of which are also of ground potential. This assumption is not supported by maintenance history or evidence found at the site.

Eyewitnesses routinely report an instantaneous ignition as an “explosion.” Even those witnesses who heard the crack or snap or clap of sudden sound could have been hearing the compromise of the center wing section as is failed under tremendous tension. At least eight witnesses on the ground reported the aircraft level and low, close over the mountains. An analysis to calculate that speed and altitude needed to be included in the original report.

Last, and important, is the consideration of the accident aircraft’s high speed and vertical acceleration factors (G-loading), and the presence of great thermal energy over the mountains to produce at least moderate turbulence over those ridges. Interestingly, one of the witnesses was piloting a glider being towed aloft at the time utilizing those very thermals.


A submission of factors to be reconsidered, a Brief

In my opinion, the loss of Tanker 82 resulted from the flexing of the wing at a speed greater than permitted by the manufacturer and the FAA (and the Forest Service), even under the most optimum environmental conditions. These conditions on August 13 th over the San Gabriel Mountains were not ideal. The 102 degree (F) heat-producing turbulence, coupled with high speed over the mountains, were the critical elements contributing to this tragic loss of crew and airplane.

The center wing section top skin, under tension and compression, gave way in catastrophic overload (seen under an 8x field microscope). It is easy to demonstrate the phenomenon by holding one’s arms out straight and flexing upward. Muscles that are aligned spanwise (stringers of the wing) will ripple. The repeated flexing of the wing skin in rapid succession in this high speed, high G-loading, both negative and positive, quickly exceeded the performance limitations. No negative acceleration greater than zero was permitted. Those negative G-forces put the wing top section into tension, followed by an up-flexing of positive G’s, put the same area into compression. The skin failed, consistent with the NTSB laboratory’s earlier finding of “vertical tensile failure,” and the low pressure present over the top of the wing, carried the pieces off, putting them at a location 2000 feet short of the fuselage impact point.

Structural integrity of the center wing section was completely compromised with the loss of its upper wing skin. The wing, consistent with some witness statements, likely separated up and back, as rapid increase in its own lift increased its own induced drag. At some point propeller cables and engine control cables were pulled, even slightly at the onset, and the number three propeller, in one instant producing about 3500 horsepower in forward, slammed into about 2200 horsepower of reverse thrust. The bulkhead holding the wing fuel over the number three engine was ripped open by the breaking wingroot at the center wing attaching point.

The witnesses who saw and heard an “explosion” were seeing the ignition of a cloud of right wing fuel as the bulkhead was compromised, which separated the infamous dry bay center wing section, from the entire right wing fuel. That right wing fuel tank is not a bladder tank, but simply a wing box filled with fuel (a wet wing). The ignition source was the hot section of the number three engine which sits directly below the location of that fuel tank bulkhead as it was compromised in the breaking of the wing.

In conclusion, the following four factors are submitted as contributing to the loss of Tanker 82 and its crew: (1) Crew-induced structural failure, (2) Aircraft performance (load) envelope exceeded, (3) Environmental (moderate turbulence); and (4) Lack of a CRM and or loss of situational awareness wherein three crewmen made a conscious decision to operate the aircraft in turbulent, high speed flight in close proximity to the ridgelines.



About the author:
Douglas R. Herlihy, is a partner of HERLIHY & LEONARD, an association of air safety investigators, and a former NTSB investigator with the Washington DC based National Go-Team. In a previous career as a Coast Guard Aviator, Commander Herlihy completed six flying tours, on the HU16 and C-130 and was assigned as Chief, Search & Rescue, for the USCG Atlantic Area. An airline transport rated pilot, with 16,400 hours pilot time, he is a member of the International Society of Airsafety Investigators, (ISASI) MO 3194, and World Investigators Network (WIN)


This report would not have been possible without the professional assistance of Daryl Shippy, A&P/IA, Mesa, Arizona; William Haggard, CCM, meteorologist, Asheville, NC; and Gary Fowler, Ph.D., metallurgist, Gardena, CA.




[1] Firefighting aircraft still operate under U.S. Federal Aviation regulations for agricultural spraying (14 C.F.R. 137) which still promotes the philosophy of single-piloted crop dusting, even though multi-piloted, multi-engine aircraft are more widely used. Part 137, has a noticeable absence of regulatory guidance in crew resource management (CRM), training, flight and duty time for crews operating under these regulations, instead leaving it to the management of the particular government agency overseeing the tankers’ contract.

[2] More than 20 Grumman S2 “Tracker” series, have been re-engined with Garrett TPE-331-14 engines, removing Wright 1820 piston engines (Marsh Aviation in Arizona).

[3] This investigators research indicates that Hemet Valley Flying Service and Aero Union Inc., may have modified as many as seven and four C130A airplanes, respectively, beginning around 1988. Only one hull, SN 56-540A was believed to have been lost.

[4] P2V “Tanker 82” operated by Neptune, Inc., for the US Forest Service, crashed at Squaw Peak near Missoula on 29 July, 1994. ( See fn. 6 ).

[56] Accident data from Aerial Firefighting Industry Association, Alexandria, VA

[7] Kelly v. USA , No. CV97-74-M-DWM, U.S. Dist. Ct. (Mont. 1997).

[8] The ex-U.S. Air Force aircraft (sn 56-540A) were modified in 1988 after desert storage with internal tanks and hydraulic doors holding 3000 gallons of retardant-water solution. Its certified maximum takeoff weight was 120,000 pounds. Limitations and operating procedures were essentially unchanged.

[9] The fuselage impacted on the southern facing ridge near elevation marked 7025 of the Pechner Canyon in the Angeles National Forest, about 5 miles southeast of Juniper Hills, California, near coordinates N34-24’-30”, W117-56’ (Geological Survey -Juniper Hills, CA quadrangle).

[10] Changes to 49 Code of Federal Regulations 830 et. seq.

[11] Accidents involving government aircraft or those contracted to government agencies, accidents were previously exempt from NTSB investigation, and had limited FAA surveillance.

[12] One eyewitness to the “explosion, ” piloting a glider under tow, was a former military jet pilot.

[13] The BLM’s Automatic Lightning Detection System (ALDS) and California Department of Forestry (CDF) discounted likelihood of lightning strike, though convective activity was strong in immediate area.

[14] A piece of wreckage was examined by a laboratory for a suspected lightning signature. That examination concluded that a small point of fused metal was likely not lightning-related.

[15] Brief of accident LAX-94-F-A-323, adopted 19 December 1995.

[16] Id. Factual Report of Accident.

[17] Investigator’s field notes “8-14-94, 11:36AM, NTSB (plus 2) dropped at crash site.”

[18] Allison turboprop T-56-A9D, 3750 SHP (shaft horsepower).

[19] The C130 aircraft originally had been fitted with 3-bladed propellers, however later modified with 4-bladed Hamilton Standard 54H60-90 propellers. Engines and propellers are numbered beginning on the left (port) outboard position, with number one mounted far left and number four far right.

[20] See photographs of a wreckage site.

[21] The aircraft’s center wing section is actually a hollow box frame 440 inches across, or extending 220 inches from the “zero” centerline. The “skin” is a fabrication of nearly _ inch thick aluminum riveted to stringers and ribs which make up the box. The forward and rear sides of the box have an “I-beam” form of construction.

[22] This aircraft had been fitted with rubber bladder type auxiliary tanks which rested in cavities between 140 and 170 inches right and left side of the centerline. Neither of these tanks was used, and may have had residual fuel inside. The auxiliary tanks and special electronics were added after manufacture in 1956-57 to extend the range for special missions for the military. They had not been removed and were found to have been filled only on one occasion since 1988.

[23] These differences, while not significant in the initial investigation, hold certain keys to understanding the findings of the re-investigation.

[24] The Allison T-56 engine is essentially divided into three distinct sections from front to back, behind the propeller and its gear-reduction box. The engine comprises of the compressor section, a 14-stage series of rotor and stationary fans, a six-chambered burner section wherein the fuel-air mixture is burned, and the last section, a four-stage power turbine through which the hot gases of nearly 1000 degrees (C) pass. Ninety percent (90%) of the power is derived from the energy imparted to the final (hot section) and is transmitted to the gearbox and the compressor through a center axial shaft. Ten percent (10%) of the energy is jet thrust through the exhaust pipe.

[25] The NTSB report stated, in part in the factual and at the “Structures Group” report, that “Very little of the cockpit instrumentation was identified.” And, “...Much of the wreckage and the impact site was covered with reddish fire retardant...No distinct soot patterns or streaking was observed on the wreckage; however, the unstable footing, severe damage to the wreckage, and inability to recover the wreckage prevented full documentation.”

[26] The on-site investigation was conducted by the NTSB on the afternoon of 14 August 1994. Absent investigators, on or about 21 August, ground personnel from the Air National Guard hooked long lines to the engines and top wing material and hoisted them from the site. A similar attempt was made to recover the cockpit, but it broke apart during the hoist attempt and was abandoned. (Interview with observer on military helicopter.)

[27] LAX-94-F-A-323 Preliminary Report.

[28] The civilian version of later C130 models, L-382, does have applicable ADs to the airframe and engines.

[29] The FAA report is titled NTSB Maintenance Records Group Field Notes, August 22, 1994. The FAA inspector was the only person listed in that group.

[30] Air Force records show the aircraft, sn 56-540A, to have been accepted for service on 11 December 1957.

[31] Inspection under 14 C.F.R.

[32] O-rings are rubber seals which fit in couplings known as “Wiggins fittings,” similar to a plumber’s pipe union, and squeezed into a seat between two compression washers and held together with a threaded and locked nut.

[33] Cannon plugs are threaded electrical plugs, wherein the electrical pin connections seat into an insulated receptacle, and locked into place by a locking nut to the plug housing, both of which are grounded to the airframe.

[34] C130A-1 manual , Emergency Section 3, pp 3-22

[35] A transmitted “squeal” at 20:31:26 GMT was likely the result of wiring, or radio units (possibly in the electronics bay below the cockpit or a force applied to them, causing an uncommanded “key-up” of the VHF-AM transceiver. (on the frequency of the Air Traffic Control -ATC- facility) Such “chirps or squeals” are not uncommon in reviewing ATC tape recordings of aircraft involved in crashes or midair collisions.

[36] ATC recording and transcript indicated an “Oh, shit” or Oh, shoot” was transmitted at 20:31:33.

[37] Turbine Inlet Temperature (T.I.T.), measured by thermocouples at the inlet stage of the power turbine behind the burner chambers.

[38] U.S. Geographic Survey, Juniper Hills Quadrangle

[39] Radar altitude recording (transmissions from “mode C” of the transponders) have tolerances of accuracy. It is not uncommon to have variations of 100 feet recorded by ATC while an aircraft is in level flight.

[40] Pallett Peak, N34-23-24, W117-54-20

[41] Depending upon mixture density, the slurry mixture can weigh between 9 and 13 pounds per gallon. Using a calculation of 10 pounds per gallon would keep the accident aircraft from a calculation exceeding the maximum takeoff weight limitation of 120,000 pounds.

[42] FAA limits because of the tanker modification reduced the allowable to 180 knots indicated airspeed (KIAS)

[43] This limit is 100% of the designed operational load factor or total force acting on the aircraft. The Ultimate Load Factor ULF are developed and tested to 150% of the load limit, to the structural extremes. However, in the words of Professor Dole (Charles E. Dole, Fundamentals of Aircraft Material Factors, USC 1989) , regarding load limits, “It should be kept in mind, however, that the tests were conducted upon new airplanes and also these airplanes undoubtedly were constructed with more care than the average production model. No corrosion or fatigue problems existed on these test airplanes. Older fleet airplanes may have reduced capability to withstand high load factors....”

[44] Used extensively to guide tankers to their target. While multi-engine, they are flown by a single pilot.

[45] See, Kelly v. USA, lack of CRM as factor in crash of P2V Tanker 4.

[46] Forest Service designation for captains to execute missions with sole authority without lead planes.