What Price Safety?
Chris Jones reporting on the ATSB
From lame’s thread “Close Calls Point To Air Safety Time Bomb” http://www.pprune.org/forums/showthr...hreadid=126418
It’s a shame that the impact of VOR’s topic has been somewhat diluted by extraneous chatter.
The ATSB’s report for last year has a chilling message for the senior and experienced aviators amongst us.
I am sure that nobody would disagree that Kaptin M's observation stated in his first line says describes the basic tenet of our profession.
Unfortunately, we also realise that while pilots have the responsibility for ensuring Safety of their passengers and aircraft, pilots are very often denied the authority to make the changes necessary to "make it so".
Last year alone there were three serious incidents – a near-miss over Tasmania, a large airliner running off the runway on landing in Darwin, and one instance where both pilots on a Boeing 737 were overcome by smoke caused by an electrical fault in the cabin.
It’s a shame that the impact of VOR’s topic has been somewhat diluted by extraneous chatter.
The ATSB’s report for last year has a chilling message for the senior and experienced aviators amongst us.
I am sure that nobody would disagree that Kaptin M's observation stated in his first line says describes the basic tenet of our profession.
Unfortunately, we also realise that while pilots have the responsibility for ensuring Safety of their passengers and aircraft, pilots are very often denied the authority to make the changes necessary to "make it so".
Thread Starter
Join Date: Jan 2004
Location: Sydney
Posts: 132
Likes: 0
Received 0 Likes
on
0 Posts
Air Traffic Control Operational Errors – Does the United States represent “Best Pract
The following is an extract of the Executive Summary from a recent [2003] report by the United States Department of Transportation – Office of the Inspector General – into Operational Errors and Runway Incursions within the United States. An “Operational Error”, in the context of this report, is effectively a serious loss of separation by air traffic controllers – one where two aircraft were within 12 seconds of collision. Note that there were 1194 reported operational errors in the United States in 2001 – i.e., 1194 instances of aircraft within 12 seconds of collision.
In the extract, though important, we have removed references to Runway Incursions – however the full report is available from http://www.oig.dot.gov/
The reason we have provided this information is two-fold. The first is to caution that to refer to the United States as a best-practice model may in fact belie the fact of your own good safety record or practices – that is, your own system may be substantially better than that in place in the United States.
If a proponent of airspace change is to be believed, the United States handles 20 times as much traffic as is handled by controllers in your country. That would seem to say that 1194/20 – or 60 very near miss incidents per year - is an acceptable level of safety for Australia. We think not.
The second is to challenge readers – and in particular your airspace designers – that as was done in the United States, evaluating the contention that you operate a safe system, and then underpinning that safety (or rectifying identified deficiencies) – is a far better use of resource than blindly implementing a new airspace system “simply because someone believes it is better”.
We apologize for the length of this extract – however we believe you should note the detailed findings and recommendations, as a caution on your own system. We are also NOT targeting air traffic controllers – as much information can be provided of pilot errors. This information is provided as a warning that safety should NEVER be taken as a given.
Operational Errors and Runway Incursions
After several years of continuous increases in operational errors and runway incursions, FAA has made progress in reducing these incidents in FY 2002. From FY 1998 through FY 2001, operational errors increased 35 percent from 885 to 1,194. This past year FAA has begun to reverse this trend. In FY 2002, operational errors decreased 11 percent to 1,061. FAA’s success was due in part to the implementation of FAA and industry initiatives. In addition, we found that there was a statistical correlation between the decrease in these incidents and the reduction in air traffic operations.
Despite FAA’s progress in reducing the number of operational errors and runway incursions in FY 2002, the number of these incidents is still too high considering the potential catastrophic results of a midair collision or a runway accident.
In FY 2002, on average, three operational errors and one runway incursion occurred each day. FAA records showed that the most serious operational errors (those rated as high risk) occurred, on average, ONCE EVERY 8 DAYS. ON AVERAGE, IN FY 2002, AT LEAST ONE COMMERCIAL AIRCRAFT WAS INVOLVED IN A SERIOUS RUNWAY INCURSION OR OPERATIONAL ERROR ONCE EVERY 10 DAYS.
In our opinion, FAA’s severity rating system understates the number of serious operational errors. For example, our analysis of operational errors that occurred from May 1, 2001, when FAA established its severity rating system, through May 31, 2002, showed that the most serious operational errors occurred, on average, once every 3 days, when we included operational errors rated moderate i.e., that were within 30 seconds of a midair collision.
In the first 5 months of FY 2003, the number of operational errors rated high risk increased from 21 to 24. While the progress in reducing serious runway incursions is encouraging, it is important that FAA take additional actions to further reduce the number and safety risk of operational errors and runway incursions, especially since FAA projects that air traffic operations will return to pre-September 11th growth patterns between 2005 and 2007.
……
FAA considers many operational errors as moderate severity when in fact some of these errors are very serious. For example, FAA rated an error as moderate that was less than 12 seconds from a midair collision. FAA needs to modify its rating system to more accurately identify the most serious operational errors, focus its resources on reducing them, and ensure that controllers receive the appropriate training for high risk errors.
FAA procedures do not require training when controllers have multiple operational errors or for controllers who have operational errors that pose a moderate or high safety risk. The procedures state only that skill enhancement training “may” be provided and, therefore, are open to interpretation.
NTSB has expressed concerns over these procedures. Under these procedures, for example, a controller had three operational errors within a 2-year period, but did not receive any training after the second and third errors because these two errors were categorized as low severity. In addition, our review of 85 moderate and high severity operational errors disclosed that the controllers involved did not receive any formal training for 18 (21 percent) of these errors. FAA needs to revise its procedures to ensure that controllers are trained to reduce the risk of future errors.
In December 2000, we reported that FAA needed to approach reducing operational errors with a sense of urgency and provide stronger national oversight to ensure efforts to reduce these errors are effective. FAA initiated actions to improve its oversight efforts and, together with the National Air Traffic Controllers Association (NATCA), established a system to identify the severity, or collision hazard, of operational errors and to focus resources on preventing the most severe errors.
The rating system uses a 100-point scale to rate the severity of errors into three risk categories - low (39 and below), moderate (40 to 89), and high (90 and above). Points are assigned based on vertical and horizontal separation distances, flight paths, closure rate, and level of air traffic control involvement (controlled: controller was aware an error was about to occur, or uncontrolled: controller was not aware an error was about to occur).
Those operational errors rated as low are not considered a safety risk and are categorized as technical violations.
The reduction in air traffic operations and FAA initiatives contributed to FAA’s progress in reducing operational errors. Most of the decrease in FY 2002 occurred at three facilities.
Specifically:
We analyzed the 11 percent decrease in operational errors from FY 2001 to FY 2002 (1,194 to 1,061) and found that the decrease was due in part to the 3 percent reduction in air traffic operations (about 4.2 million operations) during FY 2002. We performed a statistical correlation analysis between air traffic operations and operational errors on a monthly basis for FY 2001 and FY 2002 and concluded there was a positive correlation. In other words, as operations decreased, errors also decreased.
We also performed the same statistical correlation analysis for 24 facilities that had decreases in operations in FY 2002 and concluded that there was also a positive correlation on a facility basis.
There is no question that FAA has made improvements since we issued our report in December 2000; however, we found that much work remains in reducing operational errors.
For FY 2002, FAA changed its goal from reducing the total number of operational errors to reducing the most serious incidents. However, FAA missed its FY 2002 goal of having no more than 568 operational errors with less than 80 percent of required separation between aircraft with 617 such errors.
Seventy-eight percent of the 1,103 operational errors that FAA rated during the 13-month period from May 1, 2001, to May 31, 2002, posed a moderate or high safety risk.
FAA's rating system understates the safety risk of the most serious operational errors. Only 61 of 1,103 operational errors were rated high risk (90 and above out of 100 points). However, we found another 65 errors rated in the high end of moderate. (70 to 89 points) that, in our opinion, were also very serious (within 30 seconds of a midair collision).
On average, operational errors that scored between 70 and 89 points had only 50 percent of the required separation between aircraft. Depending on the closure rate of the aircraft involved, these errors can be only seconds away from an accident. For example, a controller directed two commercial airliners head-on at a closure rate of 460 miles per hour. The two aircraft had only 43 percent of the required 3-mile horizontal separation and were less than 12 seconds from a midair collision. The incident was rated as moderate scoring 87 points. In our opinion, this operational error was very serious, and FAA should consider errors such as this one to be high risk.
FAA does not identify and monitor statistics on the number of operational errors that involve commercial aircraft, general aviation aircraft, and military aircraft. This is important in measuring the overall safety impact of operational errors. For example, operational errors that involve commercial aircraft should be identified and monitored because they can place hundreds of passengers at risk.
FAA procedures do not require training when controllers have multiple operational errors or for controllers who have operational errors that posed a moderate or high safety risk. This occurred because, in 2001, FAA and NATCA signed a Memorandum of Understanding establishing new procedures for the types of actions that can be taken when operational errors occur, in part, to change the controller’s belief that actions taken as a result of an operational error were punitive. In the past, supervisors could decertify a controller and provide remedial training even if the error was the controller’s first operational error and did not pose a safety risk. We agree with FAA's decision to eliminate the perception that actions taken after an operational error are punitive.
However, we do not consider training to be punitive. As a result of FAA’s new procedures, some controllers with multiple errors or errors that posed a moderate or high safety risk did not receive training. Specifically: FAA procedures set no limit on the number of low severity errors a controller could have before training is required; therefore, a controller can have multiple errors and not receive training.
The procedures state that training will not be provided to controllers who have low severity errors that are controlled (i.e., the controller was aware the error was about to occur). For low severity errors that are uncontrolled, training “may” be provided but is not mandatory.
For example, a controller had three operational errors within a 2-year period, but did not receive any training after the second and third errors because these two errors were categorized as low severity. Less than 1 year later, the controller had a fourth operational error that was rated in the high end of moderate. While the controller did receive training after the fourth error, training provided after the second or third error possibly could have prevented the fourth error. Our review of 85 moderate and high severity operational errors disclosed that the controllers involved did not receive any formal training for 18 (21 percent) of these errors.
In the extract, though important, we have removed references to Runway Incursions – however the full report is available from http://www.oig.dot.gov/
The reason we have provided this information is two-fold. The first is to caution that to refer to the United States as a best-practice model may in fact belie the fact of your own good safety record or practices – that is, your own system may be substantially better than that in place in the United States.
If a proponent of airspace change is to be believed, the United States handles 20 times as much traffic as is handled by controllers in your country. That would seem to say that 1194/20 – or 60 very near miss incidents per year - is an acceptable level of safety for Australia. We think not.
The second is to challenge readers – and in particular your airspace designers – that as was done in the United States, evaluating the contention that you operate a safe system, and then underpinning that safety (or rectifying identified deficiencies) – is a far better use of resource than blindly implementing a new airspace system “simply because someone believes it is better”.
We apologize for the length of this extract – however we believe you should note the detailed findings and recommendations, as a caution on your own system. We are also NOT targeting air traffic controllers – as much information can be provided of pilot errors. This information is provided as a warning that safety should NEVER be taken as a given.
Operational Errors and Runway Incursions
After several years of continuous increases in operational errors and runway incursions, FAA has made progress in reducing these incidents in FY 2002. From FY 1998 through FY 2001, operational errors increased 35 percent from 885 to 1,194. This past year FAA has begun to reverse this trend. In FY 2002, operational errors decreased 11 percent to 1,061. FAA’s success was due in part to the implementation of FAA and industry initiatives. In addition, we found that there was a statistical correlation between the decrease in these incidents and the reduction in air traffic operations.
Despite FAA’s progress in reducing the number of operational errors and runway incursions in FY 2002, the number of these incidents is still too high considering the potential catastrophic results of a midair collision or a runway accident.
In FY 2002, on average, three operational errors and one runway incursion occurred each day. FAA records showed that the most serious operational errors (those rated as high risk) occurred, on average, ONCE EVERY 8 DAYS. ON AVERAGE, IN FY 2002, AT LEAST ONE COMMERCIAL AIRCRAFT WAS INVOLVED IN A SERIOUS RUNWAY INCURSION OR OPERATIONAL ERROR ONCE EVERY 10 DAYS.
In our opinion, FAA’s severity rating system understates the number of serious operational errors. For example, our analysis of operational errors that occurred from May 1, 2001, when FAA established its severity rating system, through May 31, 2002, showed that the most serious operational errors occurred, on average, once every 3 days, when we included operational errors rated moderate i.e., that were within 30 seconds of a midair collision.
In the first 5 months of FY 2003, the number of operational errors rated high risk increased from 21 to 24. While the progress in reducing serious runway incursions is encouraging, it is important that FAA take additional actions to further reduce the number and safety risk of operational errors and runway incursions, especially since FAA projects that air traffic operations will return to pre-September 11th growth patterns between 2005 and 2007.
……
FAA considers many operational errors as moderate severity when in fact some of these errors are very serious. For example, FAA rated an error as moderate that was less than 12 seconds from a midair collision. FAA needs to modify its rating system to more accurately identify the most serious operational errors, focus its resources on reducing them, and ensure that controllers receive the appropriate training for high risk errors.
FAA procedures do not require training when controllers have multiple operational errors or for controllers who have operational errors that pose a moderate or high safety risk. The procedures state only that skill enhancement training “may” be provided and, therefore, are open to interpretation.
NTSB has expressed concerns over these procedures. Under these procedures, for example, a controller had three operational errors within a 2-year period, but did not receive any training after the second and third errors because these two errors were categorized as low severity. In addition, our review of 85 moderate and high severity operational errors disclosed that the controllers involved did not receive any formal training for 18 (21 percent) of these errors. FAA needs to revise its procedures to ensure that controllers are trained to reduce the risk of future errors.
In December 2000, we reported that FAA needed to approach reducing operational errors with a sense of urgency and provide stronger national oversight to ensure efforts to reduce these errors are effective. FAA initiated actions to improve its oversight efforts and, together with the National Air Traffic Controllers Association (NATCA), established a system to identify the severity, or collision hazard, of operational errors and to focus resources on preventing the most severe errors.
The rating system uses a 100-point scale to rate the severity of errors into three risk categories - low (39 and below), moderate (40 to 89), and high (90 and above). Points are assigned based on vertical and horizontal separation distances, flight paths, closure rate, and level of air traffic control involvement (controlled: controller was aware an error was about to occur, or uncontrolled: controller was not aware an error was about to occur).
Those operational errors rated as low are not considered a safety risk and are categorized as technical violations.
The reduction in air traffic operations and FAA initiatives contributed to FAA’s progress in reducing operational errors. Most of the decrease in FY 2002 occurred at three facilities.
Specifically:
We analyzed the 11 percent decrease in operational errors from FY 2001 to FY 2002 (1,194 to 1,061) and found that the decrease was due in part to the 3 percent reduction in air traffic operations (about 4.2 million operations) during FY 2002. We performed a statistical correlation analysis between air traffic operations and operational errors on a monthly basis for FY 2001 and FY 2002 and concluded there was a positive correlation. In other words, as operations decreased, errors also decreased.
We also performed the same statistical correlation analysis for 24 facilities that had decreases in operations in FY 2002 and concluded that there was also a positive correlation on a facility basis.
There is no question that FAA has made improvements since we issued our report in December 2000; however, we found that much work remains in reducing operational errors.
For FY 2002, FAA changed its goal from reducing the total number of operational errors to reducing the most serious incidents. However, FAA missed its FY 2002 goal of having no more than 568 operational errors with less than 80 percent of required separation between aircraft with 617 such errors.
Seventy-eight percent of the 1,103 operational errors that FAA rated during the 13-month period from May 1, 2001, to May 31, 2002, posed a moderate or high safety risk.
FAA's rating system understates the safety risk of the most serious operational errors. Only 61 of 1,103 operational errors were rated high risk (90 and above out of 100 points). However, we found another 65 errors rated in the high end of moderate. (70 to 89 points) that, in our opinion, were also very serious (within 30 seconds of a midair collision).
On average, operational errors that scored between 70 and 89 points had only 50 percent of the required separation between aircraft. Depending on the closure rate of the aircraft involved, these errors can be only seconds away from an accident. For example, a controller directed two commercial airliners head-on at a closure rate of 460 miles per hour. The two aircraft had only 43 percent of the required 3-mile horizontal separation and were less than 12 seconds from a midair collision. The incident was rated as moderate scoring 87 points. In our opinion, this operational error was very serious, and FAA should consider errors such as this one to be high risk.
FAA does not identify and monitor statistics on the number of operational errors that involve commercial aircraft, general aviation aircraft, and military aircraft. This is important in measuring the overall safety impact of operational errors. For example, operational errors that involve commercial aircraft should be identified and monitored because they can place hundreds of passengers at risk.
FAA procedures do not require training when controllers have multiple operational errors or for controllers who have operational errors that posed a moderate or high safety risk. This occurred because, in 2001, FAA and NATCA signed a Memorandum of Understanding establishing new procedures for the types of actions that can be taken when operational errors occur, in part, to change the controller’s belief that actions taken as a result of an operational error were punitive. In the past, supervisors could decertify a controller and provide remedial training even if the error was the controller’s first operational error and did not pose a safety risk. We agree with FAA's decision to eliminate the perception that actions taken after an operational error are punitive.
However, we do not consider training to be punitive. As a result of FAA’s new procedures, some controllers with multiple errors or errors that posed a moderate or high safety risk did not receive training. Specifically: FAA procedures set no limit on the number of low severity errors a controller could have before training is required; therefore, a controller can have multiple errors and not receive training.
The procedures state that training will not be provided to controllers who have low severity errors that are controlled (i.e., the controller was aware the error was about to occur). For low severity errors that are uncontrolled, training “may” be provided but is not mandatory.
For example, a controller had three operational errors within a 2-year period, but did not receive any training after the second and third errors because these two errors were categorized as low severity. Less than 1 year later, the controller had a fourth operational error that was rated in the high end of moderate. While the controller did receive training after the fourth error, training provided after the second or third error possibly could have prevented the fourth error. Our review of 85 moderate and high severity operational errors disclosed that the controllers involved did not receive any formal training for 18 (21 percent) of these errors.
Thread Starter
Join Date: Jan 2004
Location: Sydney
Posts: 132
Likes: 0
Received 0 Likes
on
0 Posts
Clarification
Having re-read our previous post, we feel we should provide some clarification.
There was absolutely no intent to malign air traffic controllers - neither was there an intent to imply that controllers are in any way responsible for the safety implications associated with your current airspace reform debate - nor pilots, for that matter.
We are also very well aware that controllers cannot significantly influence the safety issues relating to your recent incidents in Class E and Class G airspace - they are a product of faulty airspace design.
Our post was meant to show that whilst certain assertions are being made about the "best practice" status of the United States airspace system, appearances can be deceptive.
Our post was meant to say that before a State embarks on a change of the nature of NAS, it had better be certain that the internal foundation upon which that change is built is robust - and that the external system on which it is modeled truly is "better" than that it replaces.
We are reasonably confident in stating that no such studies were carried out. Should an accident result from such incompetence, the architects and proponents of change might well then be judged to be criminal negligent.
There was absolutely no intent to malign air traffic controllers - neither was there an intent to imply that controllers are in any way responsible for the safety implications associated with your current airspace reform debate - nor pilots, for that matter.
We are also very well aware that controllers cannot significantly influence the safety issues relating to your recent incidents in Class E and Class G airspace - they are a product of faulty airspace design.
Our post was meant to show that whilst certain assertions are being made about the "best practice" status of the United States airspace system, appearances can be deceptive.
Our post was meant to say that before a State embarks on a change of the nature of NAS, it had better be certain that the internal foundation upon which that change is built is robust - and that the external system on which it is modeled truly is "better" than that it replaces.
We are reasonably confident in stating that no such studies were carried out. Should an accident result from such incompetence, the architects and proponents of change might well then be judged to be criminal negligent.
