US safety officials want the Federal Aviation Administration to improve the way in which crosswind data is collected and shared with pilots at airports downwind of mountainous terrain. This follows the investigation of a Boeing 737-500 overrun at Denver two years ago.

The accident happened in December 2008 to a Continental Airlines 737 departing Denver International airport's Runway 34R. The aircraft suffered severe structural damage and caught fire, but none of the 115 occupants was killed.

The National Transportation Safety Board determined the main cause was the captain not properly controlling the aircraft for the extreme and unexpected crosswind conditions during the take-off run.

But the board also stressed that when the pilots received take-off clearance they were advised of a 26kt (48km/h) crosswind, when data which was available to controllers but not typically supplied to pilots, showed the winds were stronger.

The NTSB says that a study conducted by the National Centre for Atmospheric Research showed mountain wave conditions were present at the time of the accident, resulting in strong, localised westerly winds with intermittent gusts as high as 45kt that crossed the aircraft's path during take-off ground roll.

As a result, the NTSB is recommending that the FAA conduct research into mountain wave conditions at airports subject to those circumstances, which include Anchorage, Colorado Springs, Reno and Salt Lake City in addition to Denver.

The board concludes that existing low level windshear alert systems (LLWAS) at airports could supply a better understanding of mountain wave conditions, but the ability of those systems to alert air traffic control to gusts or crosswinds could be improved.

The NTSB is urging the FAA to archive all LLWAS data obtained from airports experiencing mountain wave conditions and make the information available for use the in the potential development of an improved LLWAS algorithm for crosswind and gusty wind alerts.

The board also says that airport wind displayed on the controller's ribbon display terminal at the time of the accident would have shown 35kt winds from the west with 40kt gusts. Continental's maximum crosswind speed is 33kt during take-off and Boeing's recommendation is 40kt. But controllers did not display that information and were not required to do so, says the NTSB, which that adds no established criteria exist for controllers to supply alternate wind information to pilots.

Given those circumstances, the NTSB recommends that the FAA should require air traffic controllers at airports with multiple sources of wind information to provide pilots with the maximum wind component a pilot could encounter.

Another conclusion is that the Denver air traffic control tower's runway selection policy does not account for crosswind components when selecting a runway configuration. The board believes the FAA should require towers to develop written runway selection programmes that factor in current and developing wind conditions when considering operational advantage in runway selection.

The NTSB has sent a total of 14 recommendations to the FAA, including a recommendation that aircraft manufacturers develop type-specific, maximum crosswind take-off limitations. Other suggestions pertain to crosswind training for pilots and cockpit seat upgrades.

Source: Flight International