On the 19th of December, 2005, Chalk’s Ocean Airways flight 101 crashed into a shipping channel shortly after take-off. This fatal accident signalled the beginning of the end for Chalk’s Ocean Airways.

I wrote about the interesting history of Chalk’s Flying Service and the Grumman Mallard previously, which has a few details relevant to this incident. The main point is that Chalk’s had a fleet of sea planes, primarily Grumman Mallards, most of which had been upgraded to Grumman Mallard Turbos. At the time of the accident, they had five Mallard seaplanes.

The aircraft registration N2969 was a Grumman Mallard built in 1947 and equipped with Pratt & Whitney radial piston engines. Chalk’s purchased the Mallard in 1980 and upgraded the aircraft to a Grumman Turbo Mallard: replacing the piston engines with turbo-propeller engines and upgrading the cabin to hold up to 17 passengers. The aircraft’s avionics were also upgraded. N2969 had 31,226 total flight hours and 39,743 total cycles (where a cycle is one take-off and landing).

That day, the Mallard flew a scheduled passenger flight from Fort Lauderdale to Miami Seaplane Base.

It was the first flight of the day for the captain, who was scheduled to fly seven flights (4.1 hours flight time). The flight was meant to depart Fort Lauderdale early that morning but they were held up by the weather. The chief pilot spoke to the captain while she was waiting at Fort Lauderdale and said that she was upbeat and alert. Based on her flight times, she should have been well rested.

But although she seemed fine that morning, all was not well. Four months earlier, she’d been appointed the company’s director of safety. Her husband said that her health had been affected by the stress of work demands and that, although she was loyal to the company, she had begun applying to other airlines. She was becoming physically exhausted and was not enjoying the job any longer. She also told him that she was concerned about the company’s maintenance.

The first officer had been hired by Chalk’s Ocean Airways in April of that year and completed the company’s initial operating experience on 21 November 2005, just a month before the accident flight. The flight from Fort Lauderdale to Miami Seaplane Base was his first flight with the captain. He was scheduled to fly the remaining six flights with her that day. He was described by those who saw him that day as normal, energetic and happy.

They arrived at Miami Seaplane Base without incident. Eighteen passengers boarded for flight 101 to Bimini in the Bahamas: 15 adults and three lap-held infants. The aircraft departed Miami Sea Base at 14:38 local time.

Shortly after take off, a number of people on the beach were horrified to see the right wing separate from the rest of the aircraft followed by smoke and fire. About half of the witnesses said they heard an explosion as the wing separated.

They then watched helplessly as the aircraft descended and crashed into the water close to Miami Beach. The elapsed flight time was about one minute.

Lifeguards working the beach responded first, using jet skis to get to the scene. The Miami Coast Guard launched a helicopter and began recovery efforts within 15 minutes. The water was only about 30 feet deep but the aircraft was destroyed in the crash and none of those on board survived the impact with the water.

The weight and balance for the accident flight was filed as 13,828 pounds, which was less than the maximum take-off weight of 14,000 pounds. The zero fuel weight was 12,488 pounds, which was less (but just barely) than the maximum zero fuel weight of 12,800 pounds.

The Turbo Mallard had last been flown two days before the accident for 10 flights after routine maintenance. The first officer on those flights reported that he and the captain had conducted a 15-20 minute operational check flight before beginning the passenger flights and found no issues; the aircraft flew normally that day.

Although initial media reports theorised about terrorism or sabotage, it was quickly clear that there was no trace of explosives or evidence of an outside plot.

The main wreckage was scattered across an area 200 feet square, including the fuselage, the left wing, the left engine, the landing gear and the empennage. The right wing, with the engine still attached, was found around 160 feet north west. The centre wing box structures were fractured where the wing intersects the fuselage and the fire broke out when the right wing fuel tank was breached.

Three days after the accident, the NTSB put out a press release including photographs showing fatigue marks.

NTSB RELEASES PHOTOS OF FATIGUE CRACKS FROM MONDAY’S CHALK’S OCEAN AIRWAYS CRASH IN MIAMI

The National Transportation Safety Board has released photos depicting fatigue cracks found on the separated wing of the Chalk’s Ocean Airways plane that crashed into the water near Miami Beach on Monday.

On Tuesday afternoon, the right wing of the aircraft was recovered. The wing had separated at the inboard section, at the wing/fuselage juncture. Safety Board engineers and metallurgists agree that the signatures are consistent with fatigue fractures. Portions of the wing are at the NTSB metallurgy laboratory in Washington, DC.

As part of the investigation of this accident, which claimed 20 lives, the Safety Board will extensively examine the wing and other structures in the coming days. The probable cause of this accident has not been determined.

The Grumman Mallard has a ‘wet wing’: the fuel is stored in sealed off portions of the wing itself. The spars, the principal structural members of the wing, are oriented wingtip-to-wingtip. The vertical surface of the spar is called the web and the top and bottom of each spar are called spar caps. The crosspieces of the wing, the wing ribs, are oriented forward to aft. It’s the ribs that give the wing its cambered shape.

I had to learn a bit about how wings work in order to get through this, so hopefully I’ve understood it correctly and the following makes sense. If not, please take it to pieces in the comments (my explanation, not someone’s wing!).

The centre wing box of the Mallard wing spans the left wing station to the right wing station for 125 inches (317 centimeters) in either direction from the aircraft’s centreline. This wing box structure includes the front and rear spars, the stringers, the ribs and the left and right fuel tanks. The skin, that is, the smooth outer covering, is made of sheet metal panels that make the upper and lower survaces of the wing and the wing box structure.

The stringers are also part of the structure of the wing. They are oriented parallel to the spars but are much smaller. The three stringers for the lower skin panel are known as Z-stringers because in a cross-section, they have the shape of the letter Z.

The vertical surface of each Z-stringer (the web) has two slosh holes: round openings in the vertical surface that allows fuel inside the tank area of the wing to move forward and aft, that is, to slosh back and forth. These slosh holes are located just outboard of the wing ribs.

The design of this wing meant that the sealed seams around the tank area tended to open as a result of the wing flexing in flight, which could cause the fuel to leak.

The wing structure for the accident aircraft was certified on the 1st of November, 1943 and then amended in March 1944. At the time, there were no requirements for a fatigue analysis.

In 1963, Grumman issued warnings that fuel leaks in the Mallard’s wing could be indicators of possible structural problems, however, there’s no evidence that this was ever taken into account on this aircraft or any of the others owned and operated by Chalk’s.

The centre wing box of the right wing was fractured at the point where the wing intersected the fuselage. The investigators recovered the pieces of the right wing to reconstruct it and discovered signs of fatigue across multiple elements of the wing box structures. Overstress fracture features were found on the middle Z-stringer, the forward Z-stringer and the forward spar lower spar cap.

Doublers are intended to relieve load in the structural element to which it is attached but as you can see from the photograph, in this case, they also did a pretty good job of hiding the damage that was occurring.

Once they removed the doublers used to strengthen weak areas they discovered a chord-wise skin crack (running from the front of the wing towards the rear) that was 16 inches (40 centimeters) long. In the photograph below, taken after the doubler was removed, you can see the crack at the fuel sump drain plate and then intersecting the labelled holes.

The skin crack intersected with three unfilled machined holes in the skin, which appear to have been stop drill holes. Stop drill holes are made at the tip of a fatigue crack to stop the crack from growing until a more permanent repair can be made. The drill hole reduces the stress concentration around the tip of the crack, which slows the growth of the crack. But this is a temporary measure, not a repair in itself.

They also found evidence of fatigue in the rear Z-stringer, which started at the slosh hole. There was sealant over the fractured surface. Under the sealant, around the edge of the slosh hole, there were sanding marks.

From this evidence, it was not difficult to understand that multiple attempts to repair the crack had been made, none of them effective. After several failed attempts at stopping the crack by drilling stop-holes, the crack was repeatedly repaired by affixing doublers (metal patches intended to take over the load from the damaged part, a normal repair procedure for minor skin damage), which were riveted over the cracked skin. However, the crack continued to grow, requiring longer and longer doublers to be fitted. Although the skin crack was slowly getting longer it was not thought to be anything other than a skin problem, which could be dealt with by affixing a doubler.

The unseen cracked stringer on the accident aircraft allowed the right wing to flex more during flight which increased the bending forces at the root such that the visible skin crack slowly increased in length with each subsequent flight.

On examination of the wreckage, investigators discovered that, in addition to the external doublers, internal doublers had also been affixed to the root area of both wings. However, maintenance records for these repairs were not available. Investigators also concluded that the cracked stringer that initiated the wing loss had probably failed completely some considerable amount of time prior to the accident, leading to a substantial weakening of the wing structure.

There was also, by the way, a five-inch crack on the left wing, at the front spar’s lower spar cap, along with multiple rough regions, signs of an overstress fracture that was progressing fairly rapidly.

In going over the maintenance records, investigators found major repair work that was clearly relevant. In 1992, during inspection, the aircraft was found to have structural corrosion under the skin, along with popped rivets. When they removed the skin, they found light-to-moderate corrosion on the top aft spar. Major repair work was logged at the time.

Then on the 6th of July in 2000, major repair of the rear Z-stringer was logged. However, Chalk’s Ocean Airways seem to have lost a load of maintenance records from 2000 and 2001 so they weren’t able to provide any furture information.

The rivets used in the repair work were not installed correctly. Specifically, some of the rivets had undersized shot heads and were overdriven. One rivet, which went through the skin and stringer flange at the slosh hole, was of a larger size and a different head type than specified. They also found additional undocumented maintenance work to the rivet fasteners. But most damning, they found undocumented maintenance which consisted of one external and three internal doublers being fitted at the lower right wing skin, right at the location where the wing separated from the fuselage.

The maintenance supervisor did not recall accomplishing, or instructing one of the company’s mechanics to accomplish, the doubler repair at right WS 34 on the accident airplane. He recalled seeing the repair several times during the C3 (right wing) inspection but thought that the repair “must have been done prior to Chalk’s getting the airplane.” The maintenance supervisor stated that he did not recall seeing any paperwork for this repair.

During the postaccident interview, the Safety Board showed the maintenance supervisor a picture of stop drill holes in the area of the doubler repair at right WS 34 on the accident airplane. The maintenance supervisor stated that he had not previously seen the stop drill holes and that he did not recollect instructing any of the mechanics to accomplish a stop drill repair in the area. He further stated that the repair was not accomplished using best maintenance practices.

Because these problems occurred over a long period of time, it was easy for the maintenance people to dismiss the problematic repairs as having been done by someone else. However, there were also clearly multiple opportunities to inspect the repairs and correct the damage done. If the damaged components had been identified, the wing failure could have been prevented.

But that didn’t happen. Instead, the airline continued to ignore repeated signs that the right wing was suffering from structural damage.

In the months leading up to the accident, there were many flight log maintenance record entries for fuel leaks in the right wing root area.

July 2005

• fuel leak right wing root
• fuel leaking in ….right wing root
• fuel leaking from …right aft bottom

In September 2005 there were three entries referring to fuel links in the right wing dry bay.The dry bay is inside the centre wing box, next to the wing fuel tanks, and holds electronic, hydraulic, and mechanical components. The maintenance records show that in all of these cases, the existing selant was removed and then re-applied.

But the fault continued to be reported and fixed, and reported, and fixed.

August 2005

• Fuel leak …lt & rt wing tanks
• cleaned lt & rt fuel tanks, repared as req’d and sealed … refueled… no leaks noted at this time

December 2005

• fuel leak discrepancy was noted
• rt fuel tank bottom aft corner has lead
• rt wing fuel tank bottom aft corner leak repair and seal

Although the manufacturer stated decades earlier that fuel leaks might indicate structural problems, it seems clear that the maintenance crew at Chalk’s in 2005 did not think that this applied to their aircraft.

Now normally, operators set repair thresholds, that is, there’s only a certain number of times an issue should reoccur in a given period and if that number is exceeded, further trouble-shooting is needed. But the leaks in the right fuel tank just kept getting basic repairs and more sealant.

Chalk’s continuous airworthiness maintenance programme included CASS, Continuing Analysis and Surveillance System, which included collecting and analysing operational data. They held monthly CASS meetings to review the previous month’s maintenance discrepencies. The September 2005 meeting minutes noted that the accident aircraft’s right wing dry bay area had suffered from a fuel leak for three consecutive days. The repair method was to remove and replace the sealant in the fuel tank.

None of the CASS meeting minutes from 2005 reference the wing box structural fractures.

On October 7, 2005, Chalk’s Ocean Airways’ PMI conducted the company’s aging airplane inspection and records review. At the time, the accident airplane had accumulated 31,012 flight hours and 39,404 flight cycles. The review included the April 13, 1992, major repair to the rear spar upper spar cap at right WS 34 and the May 6, 1992, major repair to the lower wing skin at left WS 34. The review found that no further action was required for these repairs. The review did not include the doubler repair that was observed on the chordwise skin crack just outboard of right WS 34 and for which Chalk’s Ocean Airways did not have records. On November 4, 2005, the PMI sent a letter to Chalk’s Ocean Airways indicating that the aging airplane inspection and records review had been completed.

And so that was that. The aircraft passed its review without incident. No further inspections were required.

Chalk’s Ocean Airways maintenance did not carry out supplemental inspections and procedures (in addition to the Ageing Airplane Inspection and records review) because the G-73T aircraft were exempt. The supplemental inspection requirements applied only to transport-category, turbine-powered aircraft which were type-certificated after 1 January 1958, with a maximum passenger seating capacity of 30 or more seats or a maximum payload of 7,500 pounds. (All aircraft operated entirely within the state of Alaska were also exempt, which is interesting although clearly not relevant here).

3.2 Probable Cause

The National Transportation Safety Board determines that the probable cause of this accident was the in-flight failure and separation of the right wing during normal flight, which resulted from (1) the failure of the Chalk’s Ocean Airways maintenance program to identify and properly repair fatigue cracks in the right wing and (2) the failure of the Federal Aviation Administration to detect and correct deficiencies in the company’s maintenance program.

After the accident, the Safety Board wrote a letter to the FAA, expressing concern that certain aircraft, including the Grumman Mallard Turbo, were allowed to carry passengers for scheduled flights without having to undergo these supplemental inspections. The Safety Recommendation asks the FAA to require records reviews, Aging Airplane Inspections and supplemental inspections for all aircraft operated under 14 Code of Federal Regulations (CFR) Part 121 (as well as Part 129 and 135 – basically, make it a requirement for commercial services regardless of the size and age of the aircraft).

The FAA argued that aircraft which were type-certified before 1958 and whose maximum passenger seating capacity of less than 30 seats were intentionally not included because it would be cost prohibitive to do so. They estimated that by 2010, fewer than 80 aircraft would be in scheduled operation (2010 was the compliance date for supplemental inspections).

The Safety Board has classified this Safety Recommendation as “Open–Unacceptable Response.”

The board member statement attached to the report makes it clear why they felt stricter oversight was justified:

While this accident was primarily the structural failure of a very old airplane, the facts surrounding the accident and operation of Chalk’s shine a light on an important underlying consideration for any accident involving a financially strapped operator, that of economic fitness. In the past several years, we have watched several major airlines slide into bankruptcy protection, while others closed their hangar doors permanently. It has been clear that those airlines which continue to operate under bankruptcy protection are carefully scrutinized by FAA to ensure that their financial straits do not lead to a dangerous reduction of operational and maintenance safety measures. My concern about Chalk’s and other small, financially challenged small operators is two-fold: 1) that their operaational and maintenance practices receive a heightened degree of scrutiny by FAA just like the large carriers, and 2) that the added scrutiny begins during that delicate time before a company declares bankruptcy, when its managers may begin taking dangerous risks in a last-ditch effort to stay out of bankruptcy.

…

In 1988, the Safety Board recommended that FAA issue guidance to POIs to assist them in recognizing when a carrier was experiencing financial distress, meriting increased surveillance. In a positive response to the Safety Board’s recommendation, FAA issued guidance to its inspectors about reviewing the state of an airline that is in financial distress. The guidance is clear and comprehensive.

…

Unfortunately, and for no reason our investigators could discern, the guidance was not applied to Chalk’s and the operator did not receive heightened safety scrutiny, despite significant concerns stated by DOT [the Department of Transportation] about the operator’s financial condition.

The FAA did respond to the NTSB report by issuing Emergency Airworthiness Directive 2006-01-51. This instructed all operators of Grumman Mallard and Grumman Mallard Turbo aircraft to perform detailed visual inspection to detect repairs, cracking or corrosion of the wings and to remove any repair that was found to allow for inspection of the wing structure underneath the repairs. Further, to remove the the sealant from the interior of the wet bays to allow for inspection of the skins, stringers and both spars.

The visual inspection was “to be performed in accordance with a method that has been approved by the FAA.”

But the Airworthiness Directive was released without any method having been approved by the FAA. Effectively, all all Grumman Mallard and Grumman Mallard Turbo aircraft were grounded. Chalk’s Grumman fleet could no longer fly.

At the time, there 31 Mallards registered in the US, of which 15 were considered airworthy. Since then, the FAA has approved eight privately owned aircraft on a case-by-case basis to fly for an interim period of 200 flight hours. As far as I know, the remainder of the Grumman Mallards (Turbo or not) remain prohibited from further flight.

Chalk’s briefly managed to continue offering flights by leasing aircraft from another airline. However, on the 23rd of October, 2007, Chalk’s Airlines lost its operating licence. They had not paid their lease at Fort Lauderdale Airport for over six months, owing more than $11,000 in rent and landing fees. The Department of Transportation argued that they had had over a year to petition the FAA to approve their seaplanes or to acquire new planes, but as the operator had not done either, the Department of Transportation revoked its flying charter.

The full accident report is on the NTSB site and can currently be found here: In-flight Separation of Right Wing, Flying Boat, Inc. (doing business as Chalk’s Ocean Airways) Flight 101 – Grumman Turbo Mallard (G-73T), N2969 – Port of Miami, Florida – December 19, 2005