Is This the Most Dangerous Airport in America?

American jetliner attempts to parallel park.

No, this airport is probably not the most dangerous one in America. To my knowledge, there has never been an accident here. This picture, which I took about a week ago on a layover, is from Lindbergh Field in San Diego. An American Airlines A-321 is passing over the infamous parking garage structure located directly off the end of Runway 27, which is the primary landing runway.

So how do I know that Lindberg is not the most dangerous airport? After all, the picture is not an illusion. The airplanes really pass about a hundred feet or so above the parking structure. But even though the structure is really close to arriving aircraft, no plane has ever hit it. It just looks dangerous both from the ground and also from the cockpit.

But just because no airplane has hit the building, does that make it safe? And how does one even define the word safe anyway?

There's an old adage that states "Safety is no accident". It's a double entendre meaning either that the definition of safety is the state of not experiencing any accidents, or that the existence of a safe environment is not one that occurs by happenstance. The second of these definitions is the true one. The first one is false. Let me explain.

Low-Probability High-Consequence Risk Analysis

So let's say that you'd like to do a study of road intersections in your town to determine which one is the most dangerous. One way you might go about it is to head down to the local police station to research accident reports. The intersection that had the consistently highest number of accidents is your likely candidate.

But what if there had only been one accident in your town in the previous year? Unlikely as it seems, could you then state with certainty that the intersection at which it occurred is the most dangerous? It's easy to see that sample size is important in determining the usefulness of accident statistics.

When the sample size of accidents approaches zero, as is currently the case in commercial air travel, accident statistics become useless for the purposes of predicting how or where accidents will occur. Other tools are then needed to identify which practices, policies and procedures have inherent risk, the size of that risk, how much risk is acceptable and how to mitigate that which isn't. This field has become to be known as low-probability high consequence (LPHC) risk analysis.

Commercial aviation shares the problem of low probability yet high consequence risk with other large scale applications such as maritime safety, chemical and petroleum refining and nuclear power. The absolute numbers of accidents in all of these fields are very low, but the consequences of any accident can be devastating in terms of life and property. How can operators determine if they are dancing close to the edge of a disaster given the already low numbers of accidents?

The development of LPHC risk analysis was the result of the realization that even in the absence of historical data from which accident trends can be extrapolated, risk must still be identified and mitigated.

How Does It Work?

There have been many books, papers and theses written on this topic, and while a comprehensive survey is beyond the scope of this 'umble blog, several themes keep recurring as one looks around at available sources. One of those themes is that any comprehensive safety effort must involve a sector or industry wide effort to identify potential risks. 

Collaboration is essential and takes the form of industry wide safety partnerships which include both operators and regulators. One of the results of this type of effort in commercial aviation is the Aviation Safety Action Program (ASAP) in which the FAA partners with corporate and labor union representatives to focus on voluntary safety reporting while providing limited immunity from enforcement actions. This program hopes to encourage user participation and the free flow of potentially pertinent safety information.

It's essentially a get out of jail free card if you screw up and then report how it happened. There are time limits for reporting and intentional non-compliance is not covered.

Another arrow in the LPHC quiver is the collection and analysis of data from more common but less serious incidents in order to build an index of relative risk. The idea behind this effort is that any major accident usually involves a chain of events which line up or form a chain leading to a larger failure. Identifying and mitigating risk in each element in a potential chain of events reduces the likelihood that the risk in each of the individual items will be additive.

As an example of this, some years ago my airline used to fly a visual approach to Runway 15 in Burbank. This particular approach had high terrain in the area coupled with no instrument backup and terminated on a short runway with a downhill runway slope giving a false indication of being on the proper glideslope. I really enjoyed flying it because it was a challenge, but an analysis showed that many errors were being made due to the above elements coupled with the lack of experience of some crews who didn't fly it that often. Since then a decision was made to not fly the approach at night. Probably smart.

Data Driven Safety

Data collection capabilities installed on airliners using new generation digital flight data recorders (DFDRs) now allow the mass analysis of trend information on the actual operation of the aircraft. This type of data was never before available with the older analog data recorders. An industry effort called the Flight Data Analysis Program (FDAP) allows the download and use of flight data for safety purposes.

Data driven changes to training programs and even approach procedures identified as problematic have been implemented to mitigate risk which was present but not readily visible without computer analysis.

Again, going back to LA, an analysis of a particular approach into LAX showed that many airplanes were having to make steep approaches in order to land. Most pilots knew that you had to be ready to fly the approach this way, but it was the data analysis from hundreds of flights which provided the impetus to change the approach procedure to one that is more shallow.

Commercial aviation is now one of the safest methods of transportation available rivalling even elevators and orders of magnitude safer than the car you use to drive to the airport. Still, safety experts continue to use advanced data analysis and statistical methods to make it even safer.

Oh, and the most dangerous airport? That's gotta be LaGuardia...but only after you get into a cab at the airport.